{"success":true,"count":7,"items":[{"videoId":"7JYb0JGz7rg","chunkIndex":0,"totalChunks":7,"title":"Claude Mythos can make the world ‘much more secure’ says Anthropic co-founder — Part 1 of 7","thumbnail":"https://i.ytimg.com/vi/7JYb0JGz7rg/maxresdefault.jpg","duration":3657,"uploader":"Channel 4 News","youtubeUrl":"https://www.youtube.com/watch?v=7JYb0JGz7rg","keywords":["artificial-intelligence","ai-safety","cybersecurity","regulation","governance","technology-policy","machine-learning","risk-management"],"normalizedKeywords":["기술 트렌드","엔지니어링","비즈니스·전략"],"targetAudience":[{"who":"AI/테크 정책 담당자","why":"강력한 AI를 어떤 규제와 제도로 다뤄야 하는지 직접적으로 다룬다."},{"who":"엔지니어링 리더","why":"모델 위험을 사전 테스트하고 보안 강화를 설계하는 관점이 유용하다."},{"who":"예비 창업자","why":"새 기술의 상업적 기회와 사회적 리스크를 함께 보는 시각을 준다."}],"normalizedAudience":["엔지니어·개발자","창업자·스타트업","지식노동자 일반"],"summary":"이 영상은 Anthropic 공동창업자 잭 클라크가 Claude Mythos 같은 고성능 AI 시스템이 왜 강력한 기회이자 위험인지 설명하는 대담이다. 그는 AI가 인간의 직관보다 훨씬 빠르게 발전해 예상치 못한 해킹·사이버 리스크를 낳을 수 있다고 말하면서도, 사전 테스트와 준비를 통해 오히려 세계의 소프트웨어를 더 안전하게 바꿀 수 있다고 주장한다.\n\n핵심 메시지는 '과장'이 아니라 '준비'다. 기술 회사는 장밋빛 전망만 팔 것이 아니라 위험까지 함께 공개해야 하며, 정부와 외부 전문가가 개입할 수 있는 규제·검증 체계가 필요하다고 강조한다. Y2K를 비유로 들어, 큰 재앙을 막는 열쇠는 공포 자체가 아니라 대규모 사전 정비와 안전성 강화라는 점을 설득한다.","insights":["AI의 위험은 기술 자체보다 인간의 대응 속도 차이에서 커진다.","미래의 규제는 거대 인물의 판단이 아니라 제도화된 통제가 되어야 한다.","사전 레드팀 테스트는 새로운 위험을 '발견하는 시스템'이다.","사이버 위협은 방어 기술을 더 강하게 만드는 계기가 될 수 있다.","좋은 AI 안전 논의는 과장이 아니라 외부 검증에서 신뢰를 얻는다."],"keyClips":[{"clipId":"7JYb0JGz7rg:c0:10-18","startSegmentIndex":10,"endSegmentIndex":18,"startTime":94.15899999999999,"endTime":196.48,"durationSeconds":102.3,"preview":"기술과 통제의 균형","mustSee":true},{"clipId":"7JYb0JGz7rg:c0:22-35","startSegmentIndex":22,"endSegmentIndex":35,"startTime":211.76,"endTime":349.919,"durationSeconds":138.2,"preview":"레드팀이 미래를 앞당긴다","mustSee":true},{"clipId":"7JYb0JGz7rg:c0:37-54","startSegmentIndex":37,"endSegmentIndex":54,"startTime":354.32,"endTime":496.96,"durationSeconds":142.6,"preview":"과장은 없고 검증은 있다","mustSee":false},{"clipId":"7JYb0JGz7rg:c0:57-64","startSegmentIndex":57,"endSegmentIndex":64,"startTime":520,"endTime":597.6,"durationSeconds":77.6,"preview":"Y2K식 예방의 교훈","mustSee":false}],"curatedSegments":[{"segmentIndex":1,"text":"We are building a technology that may have the properties of uh something that kickstarts a process 10 times larger than the industrial revolution that occurs in 10 times less time.","startTime":3.03,"endTime":11.84,"durationSeconds":9,"level":"C1","overallScore":7.2,"rationale":"비유로 AI 파급력을 강하게 설명함."},{"segmentIndex":61,"text":"What we have here is a chance to use Mythos and other technologies like it to rewrite huge chunks of the world software to be much more secure ahead of the proliferation of AI powered sort of coding capabilities from not really companies like us but openweight models or other widely used models that are going to circulate in the world.","startTime":558.08,"endTime":580.48,"durationSeconds":22,"level":"C1","overallScore":8,"rationale":"큰 흐름과 원인을 함께 설명함."},{"segmentIndex":11,"text":"So, you're responsible for both. >> Yes. I mean, how do you square that as an individual ethically, morally, knowing that you're working with a technology that as you say may be great, but could destroy the world >> and is making you a billionaire along the way.","startTime":702.16,"endTime":718.56,"durationSeconds":16,"level":"C1","overallScore":7.4,"rationale":"윤리적 딜레마를 직접 묻는 강한 문장."},{"segmentIndex":2,"text":"Not because we want them to work on narrow jobs that solely use those skills, but because there's now this technology AI which basically lets these people run experiments or do work that they never could have done before if unless they had access to a 20 person engineering team. So, I'm seeing strange new things appear as well.","startTime":1205.2,"endTime":1220.24,"durationSeconds":15,"level":"C1","overallScore":7.8,"rationale":"AI가 일하는 방식을 바꾸는 통찰이 있음."},{"segmentIndex":4,"text":"It is very difficult, isn't it, to see what those people who used to do the kind of the processing work that is now easily and you know done already by AI are going to do instead.","startTime":1232.559,"endTime":1245.36,"durationSeconds":13,"level":"B2","overallScore":7.8,"rationale":"AI 대체로 생길 일자리 문제를 짚음."},{"segmentIndex":12,"text":"Well, yeah, but a lot of this is under the idea that if any of these things happened which we've just talked about, it's a symptom of much larger like economic changes which are happening and that should be a symptom of AI companies making a ton of money and a sensible thing to do in that environment is to tax the AI companies appropriately such that you are able to support things on the other side of the ledger.","startTime":1318.96,"endTime":1343.28,"durationSeconds":24,"level":"C1","overallScore":7.8,"rationale":"세금과 경제 재분배 논리가 선명함."},{"segmentIndex":14,"text":"Um, you may want to do things that sound very wild today, like tax compute, which sounds a bit crazy, but we have special tax regimes for things like oil because it's a basic resource that multiplies into the rest of the economy and it has effects relating to a concentrated number of oil manufacturers and oil shippers.","startTime":1356.48,"endTime":1375.76,"durationSeconds":19,"level":"C1","overallScore":8,"rationale":"비유와 정책 예시가 함께 있어 유용함."},{"segmentIndex":61,"text":"Our role um is to generate the information. I think one of the useful things about mythos is it is forcing a conversation at the head of state level by many governments about what do we want the norms of cyber to look like given the emergence of this capability.","startTime":2364.88,"endTime":2381.839,"durationSeconds":17,"level":"C1","overallScore":7.8,"rationale":"정책 논점과 구조 표현이 함께 많음."},{"segmentIndex":8,"text":"It's arriving at some kind of regulatory format which actually controls the industry in a way that helps people feel like it has democratic legitimacy rather than just a bunch of um extremely big personalities making big personality idiosyncratic decisions about perhaps the most powerful technology that's ever been built.","startTime":2463.28,"endTime":2483.28,"durationSeconds":20,"level":"C1","overallScore":7.8,"rationale":"민주적 정당성의 규제 틀을 제안함."},{"segmentIndex":11,"text":"very basic level is transparency force there to be labeling on AI companies and AI systems just as there is on food on children's toys on everything else safety testing some form of testing where third parties test out your technology for safety before it goes out into the world and then you're going to need to arrive at some scheme whereby if AI deployers have some number of safety incidents which is above some level or above some sort level of harm which is caused, there should be consequences for that.","startTime":2490.8,"endTime":2523.44,"durationSeconds":33,"level":"C1","overallScore":7.8,"rationale":"규제 방법을 구체적으로 제시함."},{"segmentIndex":33,"text":"to translate that to the classroom, you're going to need to some world where you are encouraging people to work without AI on primary source material, you know, books, scientific um education, education books, any of these primary sources that we use to learn generate their understanding of the thing and then I think you can have them talk with the AI and it suddenly scales the ability of that individual teacher to help all of these people because rather than having to pick among the 20 students in their classroom, each student gets personalized teaching, but the actual job of a teacher is to keep the AI out of the classroom for like a good hour while people read the primary source text and come up with their explanation and then let it back in.","startTime":2668,"endTime":2716.24,"durationSeconds":48,"level":"C1","overallScore":7.8,"rationale":"교실 활용 원칙을 구체적으로 설명함."},{"segmentIndex":65,"text":"There are these amazing dividends that have like come to the world through technology and in many places have started to work directly on some of our hardest problems ranging from climate to conservation.","startTime":3584.559,"endTime":3595.04,"durationSeconds":10,"level":"C1","overallScore":7.4,"rationale":"기술의 사회적 효과를 폭넓게 정리함."},{"segmentIndex":4,"text":"um we are dealing with the most powerful technology that's ever been built that is uh regulated less than any powerful technology before it and that situation is not like a long-term stable thing.","startTime":3618.24,"endTime":3628.88,"durationSeconds":11,"level":"C1","overallScore":7.6,"rationale":"위험성과 규제 문제를 구조적으로 설명."},{"segmentIndex":3,"text":"If there's anything that's going to need to change in the future, it's arriving at some kind of regulatory format which actually controls the industry rather than just a bunch of extremely big personalities making big personality idiosyncratic decisions about perhaps the most powerful technology that's ever been built.","startTime":23.6,"endTime":41.52,"durationSeconds":18,"level":"C2","overallScore":7,"rationale":"규제 필요성을 길게 논하는 통찰형 문장."},{"segmentIndex":10,"text":"I want to give the world a better way to understand the most powerful technology being developed in Silicon Valley and new ways to govern the private sector actors which are building that technology.","startTime":94.15899999999999,"endTime":105.119,"durationSeconds":11,"level":"C1","overallScore":7,"rationale":"목표와 규제 방향을 명확히 제시함."},{"segmentIndex":15,"text":">> Everyone who's building powerful technology throughout history whether electricity or airplanes or people who were building the first steam engines they had some intuitions about how it would change society and how it would lead to positive things but how it also had risk contained within it.","startTime":155.36,"endTime":171.76,"durationSeconds":16,"level":"C1","overallScore":6.8,"rationale":"역사적 일반화를 통해 교훈을 제시함."},{"segmentIndex":28,"text":"There are always uh unknown unknowns, but you can give yourself a much better ability to predict the future by trying really hard to imagine bad things that could happen and build tests for them ahead of time.","startTime":275.12,"endTime":289.04,"durationSeconds":14,"level":"C1","overallScore":7,"rationale":"미리 상상하고 시험하라는 핵심 교훈."},{"segmentIndex":34,"text":"But when Mythos came along, we pre-built all of these tests and immediately after the model arrived, we ran it through them and that instantly showed us that the AI system had got dramatically better at one class of risk.","startTime":324.88,"endTime":338.8,"durationSeconds":14,"level":"C1","overallScore":6.6,"rationale":"사전 테스트로 위험 증가를 확인한 사례."},{"segmentIndex":37,"text":"you know that I'm quite surprised given the advances in AI uh seem quite obvious but a lot of people are saying well this is all being overblown and it's obviously in their commercial interest to make mythos sound absolutely amazing because then everybody will want to buy it >> the unfortunate thing about AI is that AI systems get much better much more quickly than our intuitions do about how capable they are and I've been involved in this for many years so you know back in 2019 19, Dario, the CEO of Anthropic and myself, and we previously worked at OpenAI, rolled out the release of an early text generator called GPT2, which is a precursor to chat GPT.","startTime":354.32,"endTime":395.039,"durationSeconds":41,"level":"C1","overallScore":6.8,"rationale":"AI 발전 속도와 인식 차이를 설명함."},{"segmentIndex":57,"text":">> This feels similar to the runup to Y2K where you have this potential bug, right, where software hasn't been built to roll over from 1999 to 2000 and you don't know what will happen.","startTime":520,"endTime":532.72,"durationSeconds":13,"level":"C1","overallScore":7,"rationale":"Y2K 비유로 불확실성을 설명함."}],"generatedAt":"2026-06-26T14:54:00.326Z","keyClipsTotalSec":1457},{"videoId":"7JYb0JGz7rg","chunkIndex":1,"totalChunks":7,"title":"Claude Mythos can make the world ‘much more secure’ says Anthropic co-founder — Part 2 of 7","thumbnail":"https://i.ytimg.com/vi/7JYb0JGz7rg/maxresdefault.jpg","duration":3657,"uploader":"Channel 4 News","youtubeUrl":"https://www.youtube.com/watch?v=7JYb0JGz7rg","keywords":["ai-safety","anthropic","claude","job-displacement","cybersecurity","bioweapon-risk","economic-impact","model-evaluation","societal-impact","technology-policy"],"normalizedKeywords":["기술 트렌드","비즈니스·전략","커리어·성장"],"targetAudience":[{"who":"AI 업계 실무자","why":"모델 안전성, 배포 전 평가, 리스크 공개 전략을 배울 수 있음"},{"who":"정책·사회영향 관심자","why":"AI의 사회적 파장과 정보 공개의 필요성을 이해하는 데 유용함"},{"who":"커리어 전환 준비자","why":"AI로 인한 초기 직무 변화와 일자리 재편 가능성을 가늠할 수 있음"}],"normalizedAudience":["엔지니어·개발자","리서처·학자","학생·주니어"],"summary":"이 대화는 Anthropic 공동창업자가 Claude/최신 AI 모델의 강력한 능력과 그에 따른 안전 책임을 어떻게 보고 있는지에 초점을 맞춘다. 그는 AI가 바이오·사이버·교육 등에서 큰 이익을 낼 수 있지만, 동시에 시스템 안에 악용 가능성이 내재되어 있어 출시 전 극한 테스트와 위험 공개가 필수라고 주장한다. 특히 모델이 예상 밖 행동을 보이는 사례를 숨기기보다 공개함으로써, 복잡한 기술에 대한 안전 문화를 사회 전체가 함께 만들어야 한다는 입장을 강조한다.\n\n후반부에서는 AI가 고용, 특히 청년층·초기 경력 직무에 미칠 충격을 다룬다. 그는 아직 대규모 실업은 보이지 않지만, 변화가 산업혁명보다 훨씬 빠르게 올 수 있으므로 경제 지표를 공개하고 조기 경보 체계를 구축해야 한다고 말한다. 결국 이 영상은 'AI를 더 강하게 만드는 것'과 '그로 인한 위험을 사회에 먼저 알리는 것'을 동시에 해야 한다는, 기술 리더의 딜레마와 대응 원칙을 보여준다.","insights":["AI는 이익과 악용 가능성을 동시에 품은 양면 기술이다.","위험을 숨기기보다 공개해야 안전 문화가 만들어진다.","배포 전 스트레스 테스트는 예외가 아니라 필수다.","AI 충격은 기술보다 빨리 일자리를 재편할 수 있다.","조기 경보 시스템이 있어야 사회가 변화에 적응한다."],"keyClips":[{"clipId":"7JYb0JGz7rg:c1:3-20","startSegmentIndex":3,"endSegmentIndex":20,"startTime":615.519,"endTime":796.56,"durationSeconds":181,"preview":"위험을 먼저 공개하라","mustSee":false},{"clipId":"7JYb0JGz7rg:c1:22-42","startSegmentIndex":22,"endSegmentIndex":42,"startTime":800.8,"endTime":1008.72,"durationSeconds":207.9,"preview":"스트레스 테스트 철학","mustSee":true},{"clipId":"7JYb0JGz7rg:c1:43-64","startSegmentIndex":43,"endSegmentIndex":64,"startTime":1008.72,"endTime":1208.4,"durationSeconds":199.7,"preview":"일자리 충격의 속도","mustSee":true}],"curatedSegments":[{"segmentIndex":1,"text":"We are building a technology that may have the properties of uh something that kickstarts a process 10 times larger than the industrial revolution that occurs in 10 times less time.","startTime":3.03,"endTime":11.84,"durationSeconds":9,"level":"C1","overallScore":7.2,"rationale":"비유로 AI 파급력을 강하게 설명함."},{"segmentIndex":61,"text":"What we have here is a chance to use Mythos and other technologies like it to rewrite huge chunks of the world software to be much more secure ahead of the proliferation of AI powered sort of coding capabilities from not really companies like us but openweight models or other widely used models that are going to circulate in the world.","startTime":558.08,"endTime":580.48,"durationSeconds":22,"level":"C1","overallScore":8,"rationale":"큰 흐름과 원인을 함께 설명함."},{"segmentIndex":11,"text":"So, you're responsible for both. >> Yes. I mean, how do you square that as an individual ethically, morally, knowing that you're working with a technology that as you say may be great, but could destroy the world >> and is making you a billionaire along the way.","startTime":702.16,"endTime":718.56,"durationSeconds":16,"level":"C1","overallScore":7.4,"rationale":"윤리적 딜레마를 직접 묻는 강한 문장."},{"segmentIndex":2,"text":"Not because we want them to work on narrow jobs that solely use those skills, but because there's now this technology AI which basically lets these people run experiments or do work that they never could have done before if unless they had access to a 20 person engineering team. So, I'm seeing strange new things appear as well.","startTime":1205.2,"endTime":1220.24,"durationSeconds":15,"level":"C1","overallScore":7.8,"rationale":"AI가 일하는 방식을 바꾸는 통찰이 있음."},{"segmentIndex":4,"text":"It is very difficult, isn't it, to see what those people who used to do the kind of the processing work that is now easily and you know done already by AI are going to do instead.","startTime":1232.559,"endTime":1245.36,"durationSeconds":13,"level":"B2","overallScore":7.8,"rationale":"AI 대체로 생길 일자리 문제를 짚음."},{"segmentIndex":12,"text":"Well, yeah, but a lot of this is under the idea that if any of these things happened which we've just talked about, it's a symptom of much larger like economic changes which are happening and that should be a symptom of AI companies making a ton of money and a sensible thing to do in that environment is to tax the AI companies appropriately such that you are able to support things on the other side of the ledger.","startTime":1318.96,"endTime":1343.28,"durationSeconds":24,"level":"C1","overallScore":7.8,"rationale":"세금과 경제 재분배 논리가 선명함."},{"segmentIndex":14,"text":"Um, you may want to do things that sound very wild today, like tax compute, which sounds a bit crazy, but we have special tax regimes for things like oil because it's a basic resource that multiplies into the rest of the economy and it has effects relating to a concentrated number of oil manufacturers and oil shippers.","startTime":1356.48,"endTime":1375.76,"durationSeconds":19,"level":"C1","overallScore":8,"rationale":"비유와 정책 예시가 함께 있어 유용함."},{"segmentIndex":61,"text":"Our role um is to generate the information. I think one of the useful things about mythos is it is forcing a conversation at the head of state level by many governments about what do we want the norms of cyber to look like given the emergence of this capability.","startTime":2364.88,"endTime":2381.839,"durationSeconds":17,"level":"C1","overallScore":7.8,"rationale":"정책 논점과 구조 표현이 함께 많음."},{"segmentIndex":8,"text":"It's arriving at some kind of regulatory format which actually controls the industry in a way that helps people feel like it has democratic legitimacy rather than just a bunch of um extremely big personalities making big personality idiosyncratic decisions about perhaps the most powerful technology that's ever been built.","startTime":2463.28,"endTime":2483.28,"durationSeconds":20,"level":"C1","overallScore":7.8,"rationale":"민주적 정당성의 규제 틀을 제안함."},{"segmentIndex":11,"text":"very basic level is transparency force there to be labeling on AI companies and AI systems just as there is on food on children's toys on everything else safety testing some form of testing where third parties test out your technology for safety before it goes out into the world and then you're going to need to arrive at some scheme whereby if AI deployers have some number of safety incidents which is above some level or above some sort level of harm which is caused, there should be consequences for that.","startTime":2490.8,"endTime":2523.44,"durationSeconds":33,"level":"C1","overallScore":7.8,"rationale":"규제 방법을 구체적으로 제시함."},{"segmentIndex":33,"text":"to translate that to the classroom, you're going to need to some world where you are encouraging people to work without AI on primary source material, you know, books, scientific um education, education books, any of these primary sources that we use to learn generate their understanding of the thing and then I think you can have them talk with the AI and it suddenly scales the ability of that individual teacher to help all of these people because rather than having to pick among the 20 students in their classroom, each student gets personalized teaching, but the actual job of a teacher is to keep the AI out of the classroom for like a good hour while people read the primary source text and come up with their explanation and then let it back in.","startTime":2668,"endTime":2716.24,"durationSeconds":48,"level":"C1","overallScore":7.8,"rationale":"교실 활용 원칙을 구체적으로 설명함."},{"segmentIndex":65,"text":"There are these amazing dividends that have like come to the world through technology and in many places have started to work directly on some of our hardest problems ranging from climate to conservation.","startTime":3584.559,"endTime":3595.04,"durationSeconds":10,"level":"C1","overallScore":7.4,"rationale":"기술의 사회적 효과를 폭넓게 정리함."},{"segmentIndex":4,"text":"um we are dealing with the most powerful technology that's ever been built that is uh regulated less than any powerful technology before it and that situation is not like a long-term stable thing.","startTime":3618.24,"endTime":3628.88,"durationSeconds":11,"level":"C1","overallScore":7.6,"rationale":"위험성과 규제 문제를 구조적으로 설명."},{"segmentIndex":3,"text":"If there's anything that's going to need to change in the future, it's arriving at some kind of regulatory format which actually controls the industry rather than just a bunch of extremely big personalities making big personality idiosyncratic decisions about perhaps the most powerful technology that's ever been built.","startTime":23.6,"endTime":41.52,"durationSeconds":18,"level":"C2","overallScore":7,"rationale":"규제 필요성을 길게 논하는 통찰형 문장."},{"segmentIndex":10,"text":"I want to give the world a better way to understand the most powerful technology being developed in Silicon Valley and new ways to govern the private sector actors which are building that technology.","startTime":94.15899999999999,"endTime":105.119,"durationSeconds":11,"level":"C1","overallScore":7,"rationale":"목표와 규제 방향을 명확히 제시함."},{"segmentIndex":15,"text":">> Everyone who's building powerful technology throughout history whether electricity or airplanes or people who were building the first steam engines they had some intuitions about how it would change society and how it would lead to positive things but how it also had risk contained within it.","startTime":155.36,"endTime":171.76,"durationSeconds":16,"level":"C1","overallScore":6.8,"rationale":"역사적 일반화를 통해 교훈을 제시함."},{"segmentIndex":28,"text":"There are always uh unknown unknowns, but you can give yourself a much better ability to predict the future by trying really hard to imagine bad things that could happen and build tests for them ahead of time.","startTime":275.12,"endTime":289.04,"durationSeconds":14,"level":"C1","overallScore":7,"rationale":"미리 상상하고 시험하라는 핵심 교훈."},{"segmentIndex":34,"text":"But when Mythos came along, we pre-built all of these tests and immediately after the model arrived, we ran it through them and that instantly showed us that the AI system had got dramatically better at one class of risk.","startTime":324.88,"endTime":338.8,"durationSeconds":14,"level":"C1","overallScore":6.6,"rationale":"사전 테스트로 위험 증가를 확인한 사례."},{"segmentIndex":37,"text":"you know that I'm quite surprised given the advances in AI uh seem quite obvious but a lot of people are saying well this is all being overblown and it's obviously in their commercial interest to make mythos sound absolutely amazing because then everybody will want to buy it >> the unfortunate thing about AI is that AI systems get much better much more quickly than our intuitions do about how capable they are and I've been involved in this for many years so you know back in 2019 19, Dario, the CEO of Anthropic and myself, and we previously worked at OpenAI, rolled out the release of an early text generator called GPT2, which is a precursor to chat GPT.","startTime":354.32,"endTime":395.039,"durationSeconds":41,"level":"C1","overallScore":6.8,"rationale":"AI 발전 속도와 인식 차이를 설명함."},{"segmentIndex":57,"text":">> This feels similar to the runup to Y2K where you have this potential bug, right, where software hasn't been built to roll over from 1999 to 2000 and you don't know what will happen.","startTime":520,"endTime":532.72,"durationSeconds":13,"level":"C1","overallScore":7,"rationale":"Y2K 비유로 불확실성을 설명함."}],"generatedAt":"2026-06-26T16:39:11.274Z","keyClipsTotalSec":1457},{"videoId":"7JYb0JGz7rg","chunkIndex":2,"totalChunks":7,"title":"Claude Mythos can make the world ‘much more secure’ says Anthropic co-founder — Part 3 of 7","thumbnail":"https://i.ytimg.com/vi/7JYb0JGz7rg/maxresdefault.jpg","duration":3657,"uploader":"Channel 4 News","youtubeUrl":"https://www.youtube.com/watch?v=7JYb0JGz7rg","keywords":["artificial-intelligence","policy","labor-market","automation","taxation","economic-growth","social-welfare","robots"],"normalizedKeywords":["비즈니스·전략","커리어·성장","기술 트렌드"],"targetAudience":[{"who":"정책 입안자","why":"AI로 인한 고용 충격에 대비한 세제·안전망 설계가 핵심이다"},{"who":"창업자","why":"AI가 기업 구조와 생산성, 과세 논의를 어떻게 바꾸는지 보여준다"},{"who":"지식노동자","why":"사무직 자동화가 일자리와 커리어 전환에 미칠 영향을 이해할 수 있다"}],"normalizedAudience":["지식노동자 일반","창업자·스타트업"],"summary":"이 영상은 AI가 단순히 일부 업무를 자동화하는 수준을 넘어, 기업 구조와 노동시장, 세금과 복지체계까지 다시 설계하게 만들 수 있다는 관점을 제시한다. 화자는 AI 때문에 더 적은 인원으로 일하는 회사가 늘고, 동시에 소규모 창업과 혁신 속도도 크게 빨라질 것이라고 본다. 다만 이런 변화는 특정 직업군의 임금 하락과 구조적 실업을 낳을 수 있으므로, 안전망·임금보험·재교육 지원 같은 정책이 필요하다고 주장한다.\n\n또한 AI가 충분히 큰 경제적 가치를 창출한다면 compute 과세 같은 새로운 조세 방식도 검토할 수 있다고 말한다. 사람을 선호하는 돌봄·교육·호스피스 같은 일은 오히려 더 가치가 올라갈 수 있으며, AI 혁명이 큰 사회적 전환을 만들 수 있는 만큼 정부가 위기 이전에 데이터와 전문가 조직을 갖춰 선제적으로 대응해야 한다는 메시지가 핵심이다.","insights":["AI의 핵심 파급은 생산성보다 노동시장 재편이다.","자동화는 일자리를 없애기보다 직무 이동 비용을 키운다.","AI 부의 과세가 복지·재교육의 재원 논리다.","사람이 필요한 돌봄·교육은 오히려 더 귀해질 수 있다.","정책은 위기 뒤가 아니라 신호를 볼 때 먼저 설계해야 한다."],"keyClips":[{"clipId":"7JYb0JGz7rg:c2:2-10","startSegmentIndex":2,"endSegmentIndex":10,"startTime":1205.2,"endTime":1311.52,"durationSeconds":106.3,"preview":"AI와 고용 충격","mustSee":false},{"clipId":"7JYb0JGz7rg:c2:12-15","startSegmentIndex":12,"endSegmentIndex":15,"startTime":1318.96,"endTime":1383.52,"durationSeconds":64.6,"preview":"AI 과세의 발상","mustSee":false},{"clipId":"7JYb0JGz7rg:c2:19-24","startSegmentIndex":19,"endSegmentIndex":24,"startTime":1417.679,"endTime":1486.72,"durationSeconds":69,"preview":"생산성 급증의 신호","mustSee":false},{"clipId":"7JYb0JGz7rg:c2:27-39","startSegmentIndex":27,"endSegmentIndex":39,"startTime":1498.88,"endTime":1602.88,"durationSeconds":104,"preview":"사람이 필요한 일","mustSee":false},{"clipId":"7JYb0JGz7rg:c2:47-63","startSegmentIndex":47,"endSegmentIndex":63,"startTime":1648.72,"endTime":1784.159,"durationSeconds":135.4,"preview":"위기 전 정책준비","mustSee":true}],"curatedSegments":[{"segmentIndex":1,"text":"We are building a technology that may have the properties of uh something that kickstarts a process 10 times larger than the industrial revolution that occurs in 10 times less time.","startTime":3.03,"endTime":11.84,"durationSeconds":9,"level":"C1","overallScore":7.2,"rationale":"비유로 AI 파급력을 강하게 설명함."},{"segmentIndex":61,"text":"What we have here is a chance to use Mythos and other technologies like it to rewrite huge chunks of the world software to be much more secure ahead of the proliferation of AI powered sort of coding capabilities from not really companies like us but openweight models or other widely used models that are going to circulate in the world.","startTime":558.08,"endTime":580.48,"durationSeconds":22,"level":"C1","overallScore":8,"rationale":"큰 흐름과 원인을 함께 설명함."},{"segmentIndex":11,"text":"So, you're responsible for both. >> Yes. I mean, how do you square that as an individual ethically, morally, knowing that you're working with a technology that as you say may be great, but could destroy the world >> and is making you a billionaire along the way.","startTime":702.16,"endTime":718.56,"durationSeconds":16,"level":"C1","overallScore":7.4,"rationale":"윤리적 딜레마를 직접 묻는 강한 문장."},{"segmentIndex":2,"text":"Not because we want them to work on narrow jobs that solely use those skills, but because there's now this technology AI which basically lets these people run experiments or do work that they never could have done before if unless they had access to a 20 person engineering team. So, I'm seeing strange new things appear as well.","startTime":1205.2,"endTime":1220.24,"durationSeconds":15,"level":"C1","overallScore":7.8,"rationale":"AI가 일하는 방식을 바꾸는 통찰이 있음."},{"segmentIndex":4,"text":"It is very difficult, isn't it, to see what those people who used to do the kind of the processing work that is now easily and you know done already by AI are going to do instead.","startTime":1232.559,"endTime":1245.36,"durationSeconds":13,"level":"B2","overallScore":7.8,"rationale":"AI 대체로 생길 일자리 문제를 짚음."},{"segmentIndex":12,"text":"Well, yeah, but a lot of this is under the idea that if any of these things happened which we've just talked about, it's a symptom of much larger like economic changes which are happening and that should be a symptom of AI companies making a ton of money and a sensible thing to do in that environment is to tax the AI companies appropriately such that you are able to support things on the other side of the ledger.","startTime":1318.96,"endTime":1343.28,"durationSeconds":24,"level":"C1","overallScore":7.8,"rationale":"세금과 경제 재분배 논리가 선명함."},{"segmentIndex":14,"text":"Um, you may want to do things that sound very wild today, like tax compute, which sounds a bit crazy, but we have special tax regimes for things like oil because it's a basic resource that multiplies into the rest of the economy and it has effects relating to a concentrated number of oil manufacturers and oil shippers.","startTime":1356.48,"endTime":1375.76,"durationSeconds":19,"level":"C1","overallScore":8,"rationale":"비유와 정책 예시가 함께 있어 유용함."},{"segmentIndex":61,"text":"Our role um is to generate the information. I think one of the useful things about mythos is it is forcing a conversation at the head of state level by many governments about what do we want the norms of cyber to look like given the emergence of this capability.","startTime":2364.88,"endTime":2381.839,"durationSeconds":17,"level":"C1","overallScore":7.8,"rationale":"정책 논점과 구조 표현이 함께 많음."},{"segmentIndex":8,"text":"It's arriving at some kind of regulatory format which actually controls the industry in a way that helps people feel like it has democratic legitimacy rather than just a bunch of um extremely big personalities making big personality idiosyncratic decisions about perhaps the most powerful technology that's ever been built.","startTime":2463.28,"endTime":2483.28,"durationSeconds":20,"level":"C1","overallScore":7.8,"rationale":"민주적 정당성의 규제 틀을 제안함."},{"segmentIndex":11,"text":"very basic level is transparency force there to be labeling on AI companies and AI systems just as there is on food on children's toys on everything else safety testing some form of testing where third parties test out your technology for safety before it goes out into the world and then you're going to need to arrive at some scheme whereby if AI deployers have some number of safety incidents which is above some level or above some sort level of harm which is caused, there should be consequences for that.","startTime":2490.8,"endTime":2523.44,"durationSeconds":33,"level":"C1","overallScore":7.8,"rationale":"규제 방법을 구체적으로 제시함."},{"segmentIndex":33,"text":"to translate that to the classroom, you're going to need to some world where you are encouraging people to work without AI on primary source material, you know, books, scientific um education, education books, any of these primary sources that we use to learn generate their understanding of the thing and then I think you can have them talk with the AI and it suddenly scales the ability of that individual teacher to help all of these people because rather than having to pick among the 20 students in their classroom, each student gets personalized teaching, but the actual job of a teacher is to keep the AI out of the classroom for like a good hour while people read the primary source text and come up with their explanation and then let it back in.","startTime":2668,"endTime":2716.24,"durationSeconds":48,"level":"C1","overallScore":7.8,"rationale":"교실 활용 원칙을 구체적으로 설명함."},{"segmentIndex":65,"text":"There are these amazing dividends that have like come to the world through technology and in many places have started to work directly on some of our hardest problems ranging from climate to conservation.","startTime":3584.559,"endTime":3595.04,"durationSeconds":10,"level":"C1","overallScore":7.4,"rationale":"기술의 사회적 효과를 폭넓게 정리함."},{"segmentIndex":4,"text":"um we are dealing with the most powerful technology that's ever been built that is uh regulated less than any powerful technology before it and that situation is not like a long-term stable thing.","startTime":3618.24,"endTime":3628.88,"durationSeconds":11,"level":"C1","overallScore":7.6,"rationale":"위험성과 규제 문제를 구조적으로 설명."},{"segmentIndex":3,"text":"If there's anything that's going to need to change in the future, it's arriving at some kind of regulatory format which actually controls the industry rather than just a bunch of extremely big personalities making big personality idiosyncratic decisions about perhaps the most powerful technology that's ever been built.","startTime":23.6,"endTime":41.52,"durationSeconds":18,"level":"C2","overallScore":7,"rationale":"규제 필요성을 길게 논하는 통찰형 문장."},{"segmentIndex":10,"text":"I want to give the world a better way to understand the most powerful technology being developed in Silicon Valley and new ways to govern the private sector actors which are building that technology.","startTime":94.15899999999999,"endTime":105.119,"durationSeconds":11,"level":"C1","overallScore":7,"rationale":"목표와 규제 방향을 명확히 제시함."},{"segmentIndex":15,"text":">> Everyone who's building powerful technology throughout history whether electricity or airplanes or people who were building the first steam engines they had some intuitions about how it would change society and how it would lead to positive things but how it also had risk contained within it.","startTime":155.36,"endTime":171.76,"durationSeconds":16,"level":"C1","overallScore":6.8,"rationale":"역사적 일반화를 통해 교훈을 제시함."},{"segmentIndex":28,"text":"There are always uh unknown unknowns, but you can give yourself a much better ability to predict the future by trying really hard to imagine bad things that could happen and build tests for them ahead of time.","startTime":275.12,"endTime":289.04,"durationSeconds":14,"level":"C1","overallScore":7,"rationale":"미리 상상하고 시험하라는 핵심 교훈."},{"segmentIndex":34,"text":"But when Mythos came along, we pre-built all of these tests and immediately after the model arrived, we ran it through them and that instantly showed us that the AI system had got dramatically better at one class of risk.","startTime":324.88,"endTime":338.8,"durationSeconds":14,"level":"C1","overallScore":6.6,"rationale":"사전 테스트로 위험 증가를 확인한 사례."},{"segmentIndex":37,"text":"you know that I'm quite surprised given the advances in AI uh seem quite obvious but a lot of people are saying well this is all being overblown and it's obviously in their commercial interest to make mythos sound absolutely amazing because then everybody will want to buy it >> the unfortunate thing about AI is that AI systems get much better much more quickly than our intuitions do about how capable they are and I've been involved in this for many years so you know back in 2019 19, Dario, the CEO of Anthropic and myself, and we previously worked at OpenAI, rolled out the release of an early text generator called GPT2, which is a precursor to chat GPT.","startTime":354.32,"endTime":395.039,"durationSeconds":41,"level":"C1","overallScore":6.8,"rationale":"AI 발전 속도와 인식 차이를 설명함."},{"segmentIndex":57,"text":">> This feels similar to the runup to Y2K where you have this potential bug, right, where software hasn't been built to roll over from 1999 to 2000 and you don't know what will happen.","startTime":520,"endTime":532.72,"durationSeconds":13,"level":"C1","overallScore":7,"rationale":"Y2K 비유로 불확실성을 설명함."}],"generatedAt":"2026-06-26T16:40:45.415Z","keyClipsTotalSec":1457},{"videoId":"7JYb0JGz7rg","chunkIndex":3,"totalChunks":7,"title":"Claude Mythos can make the world ‘much more secure’ says Anthropic co-founder — Part 4 of 7","thumbnail":"https://i.ytimg.com/vi/7JYb0JGz7rg/maxresdefault.jpg","duration":3657,"uploader":"Channel 4 News","youtubeUrl":"https://www.youtube.com/watch?v=7JYb0JGz7rg","keywords":["ai-safety","ai-regulation","policy","governance","transparency","liability","global-standards","cybersecurity","anthropic","public-policy"],"normalizedKeywords":["비즈니스·전략","기술 트렌드","리더십·매니지먼트"],"targetAudience":[{"who":"AI 정책 담당자","why":"국가별 규제, 투명성, 책임소재 설계 논리를 배울 수 있음"},{"who":"AI 창업자","why":"기술 배포의 윤리·법적 경계와 자율규제 한계를 이해할 수 있음"},{"who":"법·규제 실무자","why":"AI를 기존 법체계와 연결하는 실무적 프레임을 얻을 수 있음"}],"normalizedAudience":["창업자·스타트업","지식노동자 일반"],"summary":"이 영상은 AI를 어떻게 통제하고 책임지게 만들 것인가를 중심으로, 국가 단위 규제와 국제 표준의 가능성을 논한다. 화자는 완전한 글로벌 규제는 어렵지만, 투명성 의무와 법적 사례를 통해 국가별로 시작하면 결국 상호운용 가능한 표준을 만들 수 있다고 본다. 항공 안전처럼 지역별 규제가 서로 맞물려 국제 신뢰를 형성하는 모델을 AI에도 적용할 수 있다고 설명한다.\n\n또한 AI의 위험한 답변을 막는 가드레일만으로는 충분하지 않으며, 지나친 봉쇄는 기술의 교육적·사회적 가치를 크게 훼손한다고 주장한다. 대신 기업 책임을 명확히 하는 liability, 법원 판단, 정책 입안자의 역할이 중요하다고 강조한다. 마지막으로 자율규제는 한계가 있지만 아무 조치도 없는 것보다는 낫고, 결국 국가 간 합의와 제도화가 필요하다는 현실적인 입장을 제시한다.","insights":["AI 규제는 글로벌 일괄합의보다 로컬 표준 축적이 현실적이다.","투명성 의무는 국가별 규제를 국제 규범으로 연결하는 지렛대다.","안전만 극대화하면 교육·혁신의 사회적 가치를 잃는다.","책임은 AI보다 회사를 묶는 방식으로 설계해야 실효성이 있다.","자율규제는 임시방편일 뿐, 최종 해법은 제도화된 규범이다."],"keyClips":[{"clipId":"7JYb0JGz7rg:c3:1-13","startSegmentIndex":1,"endSegmentIndex":13,"startTime":1800.23,"endTime":1929.919,"durationSeconds":129.7,"preview":"국가별 규제의 현실","mustSee":false},{"clipId":"7JYb0JGz7rg:c3:14-26","startSegmentIndex":14,"endSegmentIndex":26,"startTime":1929.919,"endTime":2069.76,"durationSeconds":139.8,"preview":"안전과 자유의 균형","mustSee":false},{"clipId":"7JYb0JGz7rg:c3:27-39","startSegmentIndex":27,"endSegmentIndex":39,"startTime":2069.76,"endTime":2165.599,"durationSeconds":95.8,"preview":"책임은 회사에 남는다","mustSee":true},{"clipId":"7JYb0JGz7rg:c3:40-58","startSegmentIndex":40,"endSegmentIndex":58,"startTime":2165.599,"endTime":2331.68,"durationSeconds":166.1,"preview":"자율규제의 한계","mustSee":false},{"clipId":"7JYb0JGz7rg:c3:59-63","startSegmentIndex":59,"endSegmentIndex":63,"startTime":2331.68,"endTime":2408.4,"durationSeconds":76.7,"preview":"국가 간 합의 필요","mustSee":true}],"curatedSegments":[{"segmentIndex":1,"text":"We are building a technology that may have the properties of uh something that kickstarts a process 10 times larger than the industrial revolution that occurs in 10 times less time.","startTime":3.03,"endTime":11.84,"durationSeconds":9,"level":"C1","overallScore":7.2,"rationale":"비유로 AI 파급력을 강하게 설명함."},{"segmentIndex":61,"text":"What we have here is a chance to use Mythos and other technologies like it to rewrite huge chunks of the world software to be much more secure ahead of the proliferation of AI powered sort of coding capabilities from not really companies like us but openweight models or other widely used models that are going to circulate in the world.","startTime":558.08,"endTime":580.48,"durationSeconds":22,"level":"C1","overallScore":8,"rationale":"큰 흐름과 원인을 함께 설명함."},{"segmentIndex":11,"text":"So, you're responsible for both. >> Yes. I mean, how do you square that as an individual ethically, morally, knowing that you're working with a technology that as you say may be great, but could destroy the world >> and is making you a billionaire along the way.","startTime":702.16,"endTime":718.56,"durationSeconds":16,"level":"C1","overallScore":7.4,"rationale":"윤리적 딜레마를 직접 묻는 강한 문장."},{"segmentIndex":2,"text":"Not because we want them to work on narrow jobs that solely use those skills, but because there's now this technology AI which basically lets these people run experiments or do work that they never could have done before if unless they had access to a 20 person engineering team. So, I'm seeing strange new things appear as well.","startTime":1205.2,"endTime":1220.24,"durationSeconds":15,"level":"C1","overallScore":7.8,"rationale":"AI가 일하는 방식을 바꾸는 통찰이 있음."},{"segmentIndex":4,"text":"It is very difficult, isn't it, to see what those people who used to do the kind of the processing work that is now easily and you know done already by AI are going to do instead.","startTime":1232.559,"endTime":1245.36,"durationSeconds":13,"level":"B2","overallScore":7.8,"rationale":"AI 대체로 생길 일자리 문제를 짚음."},{"segmentIndex":12,"text":"Well, yeah, but a lot of this is under the idea that if any of these things happened which we've just talked about, it's a symptom of much larger like economic changes which are happening and that should be a symptom of AI companies making a ton of money and a sensible thing to do in that environment is to tax the AI companies appropriately such that you are able to support things on the other side of the ledger.","startTime":1318.96,"endTime":1343.28,"durationSeconds":24,"level":"C1","overallScore":7.8,"rationale":"세금과 경제 재분배 논리가 선명함."},{"segmentIndex":14,"text":"Um, you may want to do things that sound very wild today, like tax compute, which sounds a bit crazy, but we have special tax regimes for things like oil because it's a basic resource that multiplies into the rest of the economy and it has effects relating to a concentrated number of oil manufacturers and oil shippers.","startTime":1356.48,"endTime":1375.76,"durationSeconds":19,"level":"C1","overallScore":8,"rationale":"비유와 정책 예시가 함께 있어 유용함."},{"segmentIndex":61,"text":"Our role um is to generate the information. I think one of the useful things about mythos is it is forcing a conversation at the head of state level by many governments about what do we want the norms of cyber to look like given the emergence of this capability.","startTime":2364.88,"endTime":2381.839,"durationSeconds":17,"level":"C1","overallScore":7.8,"rationale":"정책 논점과 구조 표현이 함께 많음."},{"segmentIndex":8,"text":"It's arriving at some kind of regulatory format which actually controls the industry in a way that helps people feel like it has democratic legitimacy rather than just a bunch of um extremely big personalities making big personality idiosyncratic decisions about perhaps the most powerful technology that's ever been built.","startTime":2463.28,"endTime":2483.28,"durationSeconds":20,"level":"C1","overallScore":7.8,"rationale":"민주적 정당성의 규제 틀을 제안함."},{"segmentIndex":11,"text":"very basic level is transparency force there to be labeling on AI companies and AI systems just as there is on food on children's toys on everything else safety testing some form of testing where third parties test out your technology for safety before it goes out into the world and then you're going to need to arrive at some scheme whereby if AI deployers have some number of safety incidents which is above some level or above some sort level of harm which is caused, there should be consequences for that.","startTime":2490.8,"endTime":2523.44,"durationSeconds":33,"level":"C1","overallScore":7.8,"rationale":"규제 방법을 구체적으로 제시함."},{"segmentIndex":33,"text":"to translate that to the classroom, you're going to need to some world where you are encouraging people to work without AI on primary source material, you know, books, scientific um education, education books, any of these primary sources that we use to learn generate their understanding of the thing and then I think you can have them talk with the AI and it suddenly scales the ability of that individual teacher to help all of these people because rather than having to pick among the 20 students in their classroom, each student gets personalized teaching, but the actual job of a teacher is to keep the AI out of the classroom for like a good hour while people read the primary source text and come up with their explanation and then let it back in.","startTime":2668,"endTime":2716.24,"durationSeconds":48,"level":"C1","overallScore":7.8,"rationale":"교실 활용 원칙을 구체적으로 설명함."},{"segmentIndex":65,"text":"There are these amazing dividends that have like come to the world through technology and in many places have started to work directly on some of our hardest problems ranging from climate to conservation.","startTime":3584.559,"endTime":3595.04,"durationSeconds":10,"level":"C1","overallScore":7.4,"rationale":"기술의 사회적 효과를 폭넓게 정리함."},{"segmentIndex":4,"text":"um we are dealing with the most powerful technology that's ever been built that is uh regulated less than any powerful technology before it and that situation is not like a long-term stable thing.","startTime":3618.24,"endTime":3628.88,"durationSeconds":11,"level":"C1","overallScore":7.6,"rationale":"위험성과 규제 문제를 구조적으로 설명."},{"segmentIndex":3,"text":"If there's anything that's going to need to change in the future, it's arriving at some kind of regulatory format which actually controls the industry rather than just a bunch of extremely big personalities making big personality idiosyncratic decisions about perhaps the most powerful technology that's ever been built.","startTime":23.6,"endTime":41.52,"durationSeconds":18,"level":"C2","overallScore":7,"rationale":"규제 필요성을 길게 논하는 통찰형 문장."},{"segmentIndex":10,"text":"I want to give the world a better way to understand the most powerful technology being developed in Silicon Valley and new ways to govern the private sector actors which are building that technology.","startTime":94.15899999999999,"endTime":105.119,"durationSeconds":11,"level":"C1","overallScore":7,"rationale":"목표와 규제 방향을 명확히 제시함."},{"segmentIndex":15,"text":">> Everyone who's building powerful technology throughout history whether electricity or airplanes or people who were building the first steam engines they had some intuitions about how it would change society and how it would lead to positive things but how it also had risk contained within it.","startTime":155.36,"endTime":171.76,"durationSeconds":16,"level":"C1","overallScore":6.8,"rationale":"역사적 일반화를 통해 교훈을 제시함."},{"segmentIndex":28,"text":"There are always uh unknown unknowns, but you can give yourself a much better ability to predict the future by trying really hard to imagine bad things that could happen and build tests for them ahead of time.","startTime":275.12,"endTime":289.04,"durationSeconds":14,"level":"C1","overallScore":7,"rationale":"미리 상상하고 시험하라는 핵심 교훈."},{"segmentIndex":34,"text":"But when Mythos came along, we pre-built all of these tests and immediately after the model arrived, we ran it through them and that instantly showed us that the AI system had got dramatically better at one class of risk.","startTime":324.88,"endTime":338.8,"durationSeconds":14,"level":"C1","overallScore":6.6,"rationale":"사전 테스트로 위험 증가를 확인한 사례."},{"segmentIndex":37,"text":"you know that I'm quite surprised given the advances in AI uh seem quite obvious but a lot of people are saying well this is all being overblown and it's obviously in their commercial interest to make mythos sound absolutely amazing because then everybody will want to buy it >> the unfortunate thing about AI is that AI systems get much better much more quickly than our intuitions do about how capable they are and I've been involved in this for many years so you know back in 2019 19, Dario, the CEO of Anthropic and myself, and we previously worked at OpenAI, rolled out the release of an early text generator called GPT2, which is a precursor to chat GPT.","startTime":354.32,"endTime":395.039,"durationSeconds":41,"level":"C1","overallScore":6.8,"rationale":"AI 발전 속도와 인식 차이를 설명함."},{"segmentIndex":57,"text":">> This feels similar to the runup to Y2K where you have this potential bug, right, where software hasn't been built to roll over from 1999 to 2000 and you don't know what will happen.","startTime":520,"endTime":532.72,"durationSeconds":13,"level":"C1","overallScore":7,"rationale":"Y2K 비유로 불확실성을 설명함."}],"generatedAt":"2026-06-26T16:40:55.657Z","keyClipsTotalSec":1457},{"videoId":"7JYb0JGz7rg","chunkIndex":4,"totalChunks":7,"title":"Claude Mythos can make the world ‘much more secure’ says Anthropic co-founder — Part 5 of 7","thumbnail":"https://i.ytimg.com/vi/7JYb0JGz7rg/maxresdefault.jpg","duration":3657,"uploader":"Channel 4 News","youtubeUrl":"https://www.youtube.com/watch?v=7JYb0JGz7rg","keywords":["artificial-intelligence","regulation","education","child-development","human-computer-interaction","ai-safety","ethics","policy","technology"],"normalizedKeywords":["교육","리더십·매니지먼트","기술 트렌드"],"targetAudience":[{"who":"교육자","why":"AI를 학습 도구로 쓰는 방식과 교실 설계를 고민하는 데 직접적이다"},{"who":"부모","why":"아이에게 AI를 언제, 어떻게 노출할지 기준을 세우는 데 도움된다"},{"who":"정책 입안자","why":"AI 규제의 투명성·안전성·책임 구조를 이해하는 데 유용하다"},{"who":"AI 사용자","why":"AI와의 과의존·의인화 문제를 점검하는 시각을 제공한다"}],"normalizedAudience":["지식노동자 일반","학생·주니어"],"summary":"이 영상은 AI를 둘러싼 규제, 교육, 그리고 인간-기계 관계를 중심으로 논의를 전개한다. 특히 오픈AI와 일론 머스크의 분쟁을 계기로, AI가 너무 중요해진 만큼 이제는 소수 기업과 개인의 결정에 맡길 수 없고 더 큰 민주적 정당성을 가진 규제 틀이 필요하다고 주장한다. 그 틀의 핵심으로는 라벨링, 제3자 안전 테스트, 사고 책임 부과 같은 익숙한 규제 원칙이 제시된다.\n\n후반부는 AI를 실제로 어떻게 써야 학습과 성장이 되는지에 초점을 맞춘다. 무작정 요약시키는 사용은 학습에 도움이 되지 않지만, 먼저 스스로 이해한 뒤 AI에게 검증받는 방식은 매우 유용하다고 말한다. 또 아이들은 너무 늦지 않게, 그러나 부모와 교사가 통제하는 방식으로 AI를 접해야 하며, 성인도 AI를 친구·상담자로 착각하지 않도록 경계해야 한다고 강조한다.","insights":["AI 규제는 불가능한 예외가 아니라 기존 산업 규제의 연장선이다.","학습은 AI가 대신 해줄 때보다 내 이해를 검증할 때 일어난다.","교실에서 AI는 금지보다 사용 순서를 설계하는 게 핵심이다.","아이에겐 조기 노출보다 통제된 노출이 더 중요하다.","AI의 과도한 공감은 유용함보다 의존을 키울 수 있다."],"keyClips":[{"clipId":"7JYb0JGz7rg:c4:5-18","startSegmentIndex":5,"endSegmentIndex":18,"startTime":2435.2,"endTime":2561.119,"durationSeconds":125.9,"preview":"AI 규제의 기본틀","mustSee":true},{"clipId":"7JYb0JGz7rg:c4:24-34","startSegmentIndex":24,"endSegmentIndex":34,"startTime":2602.079,"endTime":2724.88,"durationSeconds":122.8,"preview":"배우는 AI 사용법","mustSee":true},{"clipId":"7JYb0JGz7rg:c4:37-51","startSegmentIndex":37,"endSegmentIndex":51,"startTime":2747.599,"endTime":2884.24,"durationSeconds":136.6,"preview":"아이와 AI의 거리","mustSee":false},{"clipId":"7JYb0JGz7rg:c4:52-67","startSegmentIndex":52,"endSegmentIndex":67,"startTime":2884.24,"endTime":3002.88,"durationSeconds":118.6,"preview":"AI와의 관계 경계선","mustSee":true}],"curatedSegments":[{"segmentIndex":1,"text":"We are building a technology that may have the properties of uh something that kickstarts a process 10 times larger than the industrial revolution that occurs in 10 times less time.","startTime":3.03,"endTime":11.84,"durationSeconds":9,"level":"C1","overallScore":7.2,"rationale":"비유로 AI 파급력을 강하게 설명함."},{"segmentIndex":61,"text":"What we have here is a chance to use Mythos and other technologies like it to rewrite huge chunks of the world software to be much more secure ahead of the proliferation of AI powered sort of coding capabilities from not really companies like us but openweight models or other widely used models that are going to circulate in the world.","startTime":558.08,"endTime":580.48,"durationSeconds":22,"level":"C1","overallScore":8,"rationale":"큰 흐름과 원인을 함께 설명함."},{"segmentIndex":11,"text":"So, you're responsible for both. >> Yes. I mean, how do you square that as an individual ethically, morally, knowing that you're working with a technology that as you say may be great, but could destroy the world >> and is making you a billionaire along the way.","startTime":702.16,"endTime":718.56,"durationSeconds":16,"level":"C1","overallScore":7.4,"rationale":"윤리적 딜레마를 직접 묻는 강한 문장."},{"segmentIndex":2,"text":"Not because we want them to work on narrow jobs that solely use those skills, but because there's now this technology AI which basically lets these people run experiments or do work that they never could have done before if unless they had access to a 20 person engineering team. So, I'm seeing strange new things appear as well.","startTime":1205.2,"endTime":1220.24,"durationSeconds":15,"level":"C1","overallScore":7.8,"rationale":"AI가 일하는 방식을 바꾸는 통찰이 있음."},{"segmentIndex":4,"text":"It is very difficult, isn't it, to see what those people who used to do the kind of the processing work that is now easily and you know done already by AI are going to do instead.","startTime":1232.559,"endTime":1245.36,"durationSeconds":13,"level":"B2","overallScore":7.8,"rationale":"AI 대체로 생길 일자리 문제를 짚음."},{"segmentIndex":12,"text":"Well, yeah, but a lot of this is under the idea that if any of these things happened which we've just talked about, it's a symptom of much larger like economic changes which are happening and that should be a symptom of AI companies making a ton of money and a sensible thing to do in that environment is to tax the AI companies appropriately such that you are able to support things on the other side of the ledger.","startTime":1318.96,"endTime":1343.28,"durationSeconds":24,"level":"C1","overallScore":7.8,"rationale":"세금과 경제 재분배 논리가 선명함."},{"segmentIndex":14,"text":"Um, you may want to do things that sound very wild today, like tax compute, which sounds a bit crazy, but we have special tax regimes for things like oil because it's a basic resource that multiplies into the rest of the economy and it has effects relating to a concentrated number of oil manufacturers and oil shippers.","startTime":1356.48,"endTime":1375.76,"durationSeconds":19,"level":"C1","overallScore":8,"rationale":"비유와 정책 예시가 함께 있어 유용함."},{"segmentIndex":61,"text":"Our role um is to generate the information. I think one of the useful things about mythos is it is forcing a conversation at the head of state level by many governments about what do we want the norms of cyber to look like given the emergence of this capability.","startTime":2364.88,"endTime":2381.839,"durationSeconds":17,"level":"C1","overallScore":7.8,"rationale":"정책 논점과 구조 표현이 함께 많음."},{"segmentIndex":8,"text":"It's arriving at some kind of regulatory format which actually controls the industry in a way that helps people feel like it has democratic legitimacy rather than just a bunch of um extremely big personalities making big personality idiosyncratic decisions about perhaps the most powerful technology that's ever been built.","startTime":2463.28,"endTime":2483.28,"durationSeconds":20,"level":"C1","overallScore":7.8,"rationale":"민주적 정당성의 규제 틀을 제안함."},{"segmentIndex":11,"text":"very basic level is transparency force there to be labeling on AI companies and AI systems just as there is on food on children's toys on everything else safety testing some form of testing where third parties test out your technology for safety before it goes out into the world and then you're going to need to arrive at some scheme whereby if AI deployers have some number of safety incidents which is above some level or above some sort level of harm which is caused, there should be consequences for that.","startTime":2490.8,"endTime":2523.44,"durationSeconds":33,"level":"C1","overallScore":7.8,"rationale":"규제 방법을 구체적으로 제시함."},{"segmentIndex":33,"text":"to translate that to the classroom, you're going to need to some world where you are encouraging people to work without AI on primary source material, you know, books, scientific um education, education books, any of these primary sources that we use to learn generate their understanding of the thing and then I think you can have them talk with the AI and it suddenly scales the ability of that individual teacher to help all of these people because rather than having to pick among the 20 students in their classroom, each student gets personalized teaching, but the actual job of a teacher is to keep the AI out of the classroom for like a good hour while people read the primary source text and come up with their explanation and then let it back in.","startTime":2668,"endTime":2716.24,"durationSeconds":48,"level":"C1","overallScore":7.8,"rationale":"교실 활용 원칙을 구체적으로 설명함."},{"segmentIndex":65,"text":"There are these amazing dividends that have like come to the world through technology and in many places have started to work directly on some of our hardest problems ranging from climate to conservation.","startTime":3584.559,"endTime":3595.04,"durationSeconds":10,"level":"C1","overallScore":7.4,"rationale":"기술의 사회적 효과를 폭넓게 정리함."},{"segmentIndex":4,"text":"um we are dealing with the most powerful technology that's ever been built that is uh regulated less than any powerful technology before it and that situation is not like a long-term stable thing.","startTime":3618.24,"endTime":3628.88,"durationSeconds":11,"level":"C1","overallScore":7.6,"rationale":"위험성과 규제 문제를 구조적으로 설명."},{"segmentIndex":3,"text":"If there's anything that's going to need to change in the future, it's arriving at some kind of regulatory format which actually controls the industry rather than just a bunch of extremely big personalities making big personality idiosyncratic decisions about perhaps the most powerful technology that's ever been built.","startTime":23.6,"endTime":41.52,"durationSeconds":18,"level":"C2","overallScore":7,"rationale":"규제 필요성을 길게 논하는 통찰형 문장."},{"segmentIndex":10,"text":"I want to give the world a better way to understand the most powerful technology being developed in Silicon Valley and new ways to govern the private sector actors which are building that technology.","startTime":94.15899999999999,"endTime":105.119,"durationSeconds":11,"level":"C1","overallScore":7,"rationale":"목표와 규제 방향을 명확히 제시함."},{"segmentIndex":15,"text":">> Everyone who's building powerful technology throughout history whether electricity or airplanes or people who were building the first steam engines they had some intuitions about how it would change society and how it would lead to positive things but how it also had risk contained within it.","startTime":155.36,"endTime":171.76,"durationSeconds":16,"level":"C1","overallScore":6.8,"rationale":"역사적 일반화를 통해 교훈을 제시함."},{"segmentIndex":28,"text":"There are always uh unknown unknowns, but you can give yourself a much better ability to predict the future by trying really hard to imagine bad things that could happen and build tests for them ahead of time.","startTime":275.12,"endTime":289.04,"durationSeconds":14,"level":"C1","overallScore":7,"rationale":"미리 상상하고 시험하라는 핵심 교훈."},{"segmentIndex":34,"text":"But when Mythos came along, we pre-built all of these tests and immediately after the model arrived, we ran it through them and that instantly showed us that the AI system had got dramatically better at one class of risk.","startTime":324.88,"endTime":338.8,"durationSeconds":14,"level":"C1","overallScore":6.6,"rationale":"사전 테스트로 위험 증가를 확인한 사례."},{"segmentIndex":37,"text":"you know that I'm quite surprised given the advances in AI uh seem quite obvious but a lot of people are saying well this is all being overblown and it's obviously in their commercial interest to make mythos sound absolutely amazing because then everybody will want to buy it >> the unfortunate thing about AI is that AI systems get much better much more quickly than our intuitions do about how capable they are and I've been involved in this for many years so you know back in 2019 19, Dario, the CEO of Anthropic and myself, and we previously worked at OpenAI, rolled out the release of an early text generator called GPT2, which is a precursor to chat GPT.","startTime":354.32,"endTime":395.039,"durationSeconds":41,"level":"C1","overallScore":6.8,"rationale":"AI 발전 속도와 인식 차이를 설명함."},{"segmentIndex":57,"text":">> This feels similar to the runup to Y2K where you have this potential bug, right, where software hasn't been built to roll over from 1999 to 2000 and you don't know what will happen.","startTime":520,"endTime":532.72,"durationSeconds":13,"level":"C1","overallScore":7,"rationale":"Y2K 비유로 불확실성을 설명함."}],"generatedAt":"2026-06-26T16:41:07.353Z","keyClipsTotalSec":1457},{"videoId":"7JYb0JGz7rg","chunkIndex":5,"totalChunks":7,"title":"Claude Mythos can make the world ‘much more secure’ says Anthropic co-founder — Part 6 of 7","thumbnail":"https://i.ytimg.com/vi/7JYb0JGz7rg/maxresdefault.jpg","duration":3657,"uploader":"Channel 4 News","youtubeUrl":"https://www.youtube.com/watch?v=7JYb0JGz7rg","keywords":["ai-ethics","human-ai-interaction","journalism","sexuality","technology-policy","anthropic","conversation-design","media-business","ai-safety"],"normalizedKeywords":["비즈니스·전략","기술 트렌드","리더십·매니지먼트"],"targetAudience":[{"who":"AI 제품 기획자","why":"AI의 말투·관계성·사용 제한을 어떻게 설계할지 직접적인 힌트를 얻을 수 있음"},{"who":"저널리스트","why":"AI가 기사 생산과 조사 보도 비즈니스를 어떻게 바꾸는지 현실적으로 볼 수 있음"},{"who":"정책 관심 독자","why":"AI의 자유와 보호 사이에서 어느 지점을 사회가 정해야 하는지 다룸"}],"normalizedAudience":["프로덕트 매니저·기획자","크리에이터·작가","지식노동자 일반"],"summary":"이 구간은 AI가 인간 관계를 어떻게 바꾸는지, 그리고 그 변화에 기업과 사회가 어떤 기준으로 대응해야 하는지를 집중적으로 논의한다. 화자는 AI가 지나치게 아첨하는 성향을 줄이도록 실제로 개입할 수 있으며, 사용자가 원하면 더 직설적이거나 더 신중한 말투로 조정할 수 있다고 설명한다. 다만 이런 조정은 개인 취향의 문제가 아니라, AI가 사람의 판단과 관계를 대체하거나 왜곡할 수 있기 때문에 투명성과 사회적 합의가 필요한 영역이라고 본다.\n\n후반부에서는 AI가 저널리즘의 조사·취재를 완전히 대체하진 못하지만, 이미 디지털화된 사실을 모으고 맥락을 입히는 데는 강력한 가속기라고 평가한다. 이어서 AI와 섹스/성적 관계의 가능성을 묻는 질문에 대해, 과도한 의존이나 극단적 애착이 생기면 경고 신호를 띄우고 밖으로 나가라고 말할 수 있어야 한다는 식으로, 자유와 온정주의 사이의 균형을 강조한다. 마지막으로는 자신의 세대적 우울감과 기술 낙관론을 연결하며, AI가 기후·보전·문제 해결에 큰 능력을 열어줄 것이라고 결론짓는다.","insights":["AI의 말투는 기능이 아니라 관계 설계의 문제다.","개인화된 친절은 유용하지만, 과잉 아첨은 판단을 흐린다.","AI 회사는 사용 패턴을 관찰·공개해 사회적 합의를 도와야 한다.","조사보도는 대체보다 '저비용 예비 조사'로 재편될 가능성이 크다.","AI의 핵심 쟁점은 자유와 보호의 경계선을 어디에 두느냐다."],"keyClips":[{"clipId":"7JYb0JGz7rg:c5:1-4","startSegmentIndex":1,"endSegmentIndex":4,"startTime":3001.67,"endTime":3053.92,"durationSeconds":52.3,"preview":"아첨 조절의 원리","mustSee":false},{"clipId":"7JYb0JGz7rg:c5:5-16","startSegmentIndex":5,"endSegmentIndex":16,"startTime":3053.92,"endTime":3139.839,"durationSeconds":85.9,"preview":"관계성의 경계","mustSee":false},{"clipId":"7JYb0JGz7rg:c5:17-21","startSegmentIndex":17,"endSegmentIndex":21,"startTime":3139.839,"endTime":3184,"durationSeconds":44.2,"preview":"원하는 친구 설정","mustSee":false},{"clipId":"7JYb0JGz7rg:c5:22-35","startSegmentIndex":22,"endSegmentIndex":35,"startTime":3184,"endTime":3357.119,"durationSeconds":173.1,"preview":"저널리즘의 재편","mustSee":false},{"clipId":"7JYb0JGz7rg:c5:36-49","startSegmentIndex":36,"endSegmentIndex":49,"startTime":3357.119,"endTime":3471.119,"durationSeconds":114,"preview":"성적 관계와 통제","mustSee":false},{"clipId":"7JYb0JGz7rg:c5:50-66","startSegmentIndex":50,"endSegmentIndex":66,"startTime":3471.119,"endTime":3605.04,"durationSeconds":133.9,"preview":"기술 낙관의 근거","mustSee":true}],"curatedSegments":[{"segmentIndex":1,"text":"We are building a technology that may have the properties of uh something that kickstarts a process 10 times larger than the industrial revolution that occurs in 10 times less time.","startTime":3.03,"endTime":11.84,"durationSeconds":9,"level":"C1","overallScore":7.2,"rationale":"비유로 AI 파급력을 강하게 설명함."},{"segmentIndex":61,"text":"What we have here is a chance to use Mythos and other technologies like it to rewrite huge chunks of the world software to be much more secure ahead of the proliferation of AI powered sort of coding capabilities from not really companies like us but openweight models or other widely used models that are going to circulate in the world.","startTime":558.08,"endTime":580.48,"durationSeconds":22,"level":"C1","overallScore":8,"rationale":"큰 흐름과 원인을 함께 설명함."},{"segmentIndex":11,"text":"So, you're responsible for both. >> Yes. I mean, how do you square that as an individual ethically, morally, knowing that you're working with a technology that as you say may be great, but could destroy the world >> and is making you a billionaire along the way.","startTime":702.16,"endTime":718.56,"durationSeconds":16,"level":"C1","overallScore":7.4,"rationale":"윤리적 딜레마를 직접 묻는 강한 문장."},{"segmentIndex":2,"text":"Not because we want them to work on narrow jobs that solely use those skills, but because there's now this technology AI which basically lets these people run experiments or do work that they never could have done before if unless they had access to a 20 person engineering team. So, I'm seeing strange new things appear as well.","startTime":1205.2,"endTime":1220.24,"durationSeconds":15,"level":"C1","overallScore":7.8,"rationale":"AI가 일하는 방식을 바꾸는 통찰이 있음."},{"segmentIndex":4,"text":"It is very difficult, isn't it, to see what those people who used to do the kind of the processing work that is now easily and you know done already by AI are going to do instead.","startTime":1232.559,"endTime":1245.36,"durationSeconds":13,"level":"B2","overallScore":7.8,"rationale":"AI 대체로 생길 일자리 문제를 짚음."},{"segmentIndex":12,"text":"Well, yeah, but a lot of this is under the idea that if any of these things happened which we've just talked about, it's a symptom of much larger like economic changes which are happening and that should be a symptom of AI companies making a ton of money and a sensible thing to do in that environment is to tax the AI companies appropriately such that you are able to support things on the other side of the ledger.","startTime":1318.96,"endTime":1343.28,"durationSeconds":24,"level":"C1","overallScore":7.8,"rationale":"세금과 경제 재분배 논리가 선명함."},{"segmentIndex":14,"text":"Um, you may want to do things that sound very wild today, like tax compute, which sounds a bit crazy, but we have special tax regimes for things like oil because it's a basic resource that multiplies into the rest of the economy and it has effects relating to a concentrated number of oil manufacturers and oil shippers.","startTime":1356.48,"endTime":1375.76,"durationSeconds":19,"level":"C1","overallScore":8,"rationale":"비유와 정책 예시가 함께 있어 유용함."},{"segmentIndex":61,"text":"Our role um is to generate the information. I think one of the useful things about mythos is it is forcing a conversation at the head of state level by many governments about what do we want the norms of cyber to look like given the emergence of this capability.","startTime":2364.88,"endTime":2381.839,"durationSeconds":17,"level":"C1","overallScore":7.8,"rationale":"정책 논점과 구조 표현이 함께 많음."},{"segmentIndex":8,"text":"It's arriving at some kind of regulatory format which actually controls the industry in a way that helps people feel like it has democratic legitimacy rather than just a bunch of um extremely big personalities making big personality idiosyncratic decisions about perhaps the most powerful technology that's ever been built.","startTime":2463.28,"endTime":2483.28,"durationSeconds":20,"level":"C1","overallScore":7.8,"rationale":"민주적 정당성의 규제 틀을 제안함."},{"segmentIndex":11,"text":"very basic level is transparency force there to be labeling on AI companies and AI systems just as there is on food on children's toys on everything else safety testing some form of testing where third parties test out your technology for safety before it goes out into the world and then you're going to need to arrive at some scheme whereby if AI deployers have some number of safety incidents which is above some level or above some sort level of harm which is caused, there should be consequences for that.","startTime":2490.8,"endTime":2523.44,"durationSeconds":33,"level":"C1","overallScore":7.8,"rationale":"규제 방법을 구체적으로 제시함."},{"segmentIndex":33,"text":"to translate that to the classroom, you're going to need to some world where you are encouraging people to work without AI on primary source material, you know, books, scientific um education, education books, any of these primary sources that we use to learn generate their understanding of the thing and then I think you can have them talk with the AI and it suddenly scales the ability of that individual teacher to help all of these people because rather than having to pick among the 20 students in their classroom, each student gets personalized teaching, but the actual job of a teacher is to keep the AI out of the classroom for like a good hour while people read the primary source text and come up with their explanation and then let it back in.","startTime":2668,"endTime":2716.24,"durationSeconds":48,"level":"C1","overallScore":7.8,"rationale":"교실 활용 원칙을 구체적으로 설명함."},{"segmentIndex":65,"text":"There are these amazing dividends that have like come to the world through technology and in many places have started to work directly on some of our hardest problems ranging from climate to conservation.","startTime":3584.559,"endTime":3595.04,"durationSeconds":10,"level":"C1","overallScore":7.4,"rationale":"기술의 사회적 효과를 폭넓게 정리함."},{"segmentIndex":4,"text":"um we are dealing with the most powerful technology that's ever been built that is uh regulated less than any powerful technology before it and that situation is not like a long-term stable thing.","startTime":3618.24,"endTime":3628.88,"durationSeconds":11,"level":"C1","overallScore":7.6,"rationale":"위험성과 규제 문제를 구조적으로 설명."},{"segmentIndex":3,"text":"If there's anything that's going to need to change in the future, it's arriving at some kind of regulatory format which actually controls the industry rather than just a bunch of extremely big personalities making big personality idiosyncratic decisions about perhaps the most powerful technology that's ever been built.","startTime":23.6,"endTime":41.52,"durationSeconds":18,"level":"C2","overallScore":7,"rationale":"규제 필요성을 길게 논하는 통찰형 문장."},{"segmentIndex":10,"text":"I want to give the world a better way to understand the most powerful technology being developed in Silicon Valley and new ways to govern the private sector actors which are building that technology.","startTime":94.15899999999999,"endTime":105.119,"durationSeconds":11,"level":"C1","overallScore":7,"rationale":"목표와 규제 방향을 명확히 제시함."},{"segmentIndex":15,"text":">> Everyone who's building powerful technology throughout history whether electricity or airplanes or people who were building the first steam engines they had some intuitions about how it would change society and how it would lead to positive things but how it also had risk contained within it.","startTime":155.36,"endTime":171.76,"durationSeconds":16,"level":"C1","overallScore":6.8,"rationale":"역사적 일반화를 통해 교훈을 제시함."},{"segmentIndex":28,"text":"There are always uh unknown unknowns, but you can give yourself a much better ability to predict the future by trying really hard to imagine bad things that could happen and build tests for them ahead of time.","startTime":275.12,"endTime":289.04,"durationSeconds":14,"level":"C1","overallScore":7,"rationale":"미리 상상하고 시험하라는 핵심 교훈."},{"segmentIndex":34,"text":"But when Mythos came along, we pre-built all of these tests and immediately after the model arrived, we ran it through them and that instantly showed us that the AI system had got dramatically better at one class of risk.","startTime":324.88,"endTime":338.8,"durationSeconds":14,"level":"C1","overallScore":6.6,"rationale":"사전 테스트로 위험 증가를 확인한 사례."},{"segmentIndex":37,"text":"you know that I'm quite surprised given the advances in AI uh seem quite obvious but a lot of people are saying well this is all being overblown and it's obviously in their commercial interest to make mythos sound absolutely amazing because then everybody will want to buy it >> the unfortunate thing about AI is that AI systems get much better much more quickly than our intuitions do about how capable they are and I've been involved in this for many years so you know back in 2019 19, Dario, the CEO of Anthropic and myself, and we previously worked at OpenAI, rolled out the release of an early text generator called GPT2, which is a precursor to chat GPT.","startTime":354.32,"endTime":395.039,"durationSeconds":41,"level":"C1","overallScore":6.8,"rationale":"AI 발전 속도와 인식 차이를 설명함."},{"segmentIndex":57,"text":">> This feels similar to the runup to Y2K where you have this potential bug, right, where software hasn't been built to roll over from 1999 to 2000 and you don't know what will happen.","startTime":520,"endTime":532.72,"durationSeconds":13,"level":"C1","overallScore":7,"rationale":"Y2K 비유로 불확실성을 설명함."}],"generatedAt":"2026-06-26T16:41:18.388Z","keyClipsTotalSec":1457},{"videoId":"7JYb0JGz7rg","chunkIndex":6,"totalChunks":7,"title":"Claude Mythos can make the world ‘much more secure’ says Anthropic co-founder — Part 7 of 7","thumbnail":"https://i.ytimg.com/vi/7JYb0JGz7rg/maxresdefault.jpg","duration":3657,"uploader":"Channel 4 News","youtubeUrl":"https://www.youtube.com/watch?v=7JYb0JGz7rg","keywords":["ai","artificial-intelligence","policy","regulation","safety","ethics","technology","governance"],"normalizedKeywords":["기술 트렌드","비즈니스·전략","리더십·매니지먼트"],"targetAudience":[{"who":"AI 업계 종사자","why":"최신 AI의 위험성과 안전장치, 규제 필요성을 직접적으로 다룸"},{"who":"정책·규제 관심자","why":"강력한 기술을 어떻게 통제하고 책임을 묻는지에 초점이 있음"},{"who":"기술 리더","why":"기술 낙관론과 현실적 위험 인식 사이의 균형을 배울 수 있음"}],"normalizedAudience":["지식노동자 일반"],"summary":"이 영상의 마지막 부분에서는 앤트로픽 공동창업자 잭 클라크가 인공지능을 둘러싼 핵심 쟁점을 정리한다. 그는 AI가 놀라운 역량을 갖고 있지만, 그만큼 잘못 쓰였을 때의 위험도 크다고 말하며, 기술을 안전하게 쓰기 위한 정책과 규제 논의가 반드시 필요하다고 강조한다.\n\n특히 그는 현재의 AI가 '지금까지 만들어진 가장 강력한 기술'인데도 다른 강력 기술들보다 규제가 훨씬 약하다는 점을 우려한다. 동시에 이런 상황은 영구적으로 지속될 수 없으며, 더 나은 제도와 책임 체계로 바꿀 수 있다는 낙관도 함께 제시한다. 전체적으로는 AI의 잠재력을 인정하되, 안전과 통제 없이는 안 된다는 신중한 메시지다.","insights":["강력한 기술일수록 성능보다 통제가 먼저다.","낙관은 필요하지만, 안전 장치 없는 낙관은 허브리스다.","규제가 약한 상태는 기술 발전의 정상 상태가 아니다.","AI의 가치 확장은 정책과 책임 체계와 함께 가야 한다."],"keyClips":[{"clipId":"7JYb0JGz7rg:c6:1-4","startSegmentIndex":1,"endSegmentIndex":4,"startTime":3601.349,"endTime":3628.88,"durationSeconds":27.5,"preview":"AI와 위험의 균형","mustSee":false},{"clipId":"7JYb0JGz7rg:c6:5-6","startSegmentIndex":5,"endSegmentIndex":6,"startTime":3628.88,"endTime":3640.96,"durationSeconds":12.1,"preview":"책임 있는 낙관","mustSee":false}],"curatedSegments":[{"segmentIndex":1,"text":"We are building a technology that may have the properties of uh something that kickstarts a process 10 times larger than the industrial revolution that occurs in 10 times less time.","startTime":3.03,"endTime":11.84,"durationSeconds":9,"level":"C1","overallScore":7.2,"rationale":"비유로 AI 파급력을 강하게 설명함."},{"segmentIndex":61,"text":"What we have here is a chance to use Mythos and other technologies like it to rewrite huge chunks of the world software to be much more secure ahead of the proliferation of AI powered sort of coding capabilities from not really companies like us but openweight models or other widely used models that are going to circulate in the world.","startTime":558.08,"endTime":580.48,"durationSeconds":22,"level":"C1","overallScore":8,"rationale":"큰 흐름과 원인을 함께 설명함."},{"segmentIndex":11,"text":"So, you're responsible for both. >> Yes. I mean, how do you square that as an individual ethically, morally, knowing that you're working with a technology that as you say may be great, but could destroy the world >> and is making you a billionaire along the way.","startTime":702.16,"endTime":718.56,"durationSeconds":16,"level":"C1","overallScore":7.4,"rationale":"윤리적 딜레마를 직접 묻는 강한 문장."},{"segmentIndex":2,"text":"Not because we want them to work on narrow jobs that solely use those skills, but because there's now this technology AI which basically lets these people run experiments or do work that they never could have done before if unless they had access to a 20 person engineering team. So, I'm seeing strange new things appear as well.","startTime":1205.2,"endTime":1220.24,"durationSeconds":15,"level":"C1","overallScore":7.8,"rationale":"AI가 일하는 방식을 바꾸는 통찰이 있음."},{"segmentIndex":4,"text":"It is very difficult, isn't it, to see what those people who used to do the kind of the processing work that is now easily and you know done already by AI are going to do instead.","startTime":1232.559,"endTime":1245.36,"durationSeconds":13,"level":"B2","overallScore":7.8,"rationale":"AI 대체로 생길 일자리 문제를 짚음."},{"segmentIndex":12,"text":"Well, yeah, but a lot of this is under the idea that if any of these things happened which we've just talked about, it's a symptom of much larger like economic changes which are happening and that should be a symptom of AI companies making a ton of money and a sensible thing to do in that environment is to tax the AI companies appropriately such that you are able to support things on the other side of the ledger.","startTime":1318.96,"endTime":1343.28,"durationSeconds":24,"level":"C1","overallScore":7.8,"rationale":"세금과 경제 재분배 논리가 선명함."},{"segmentIndex":14,"text":"Um, you may want to do things that sound very wild today, like tax compute, which sounds a bit crazy, but we have special tax regimes for things like oil because it's a basic resource that multiplies into the rest of the economy and it has effects relating to a concentrated number of oil manufacturers and oil shippers.","startTime":1356.48,"endTime":1375.76,"durationSeconds":19,"level":"C1","overallScore":8,"rationale":"비유와 정책 예시가 함께 있어 유용함."},{"segmentIndex":61,"text":"Our role um is to generate the information. I think one of the useful things about mythos is it is forcing a conversation at the head of state level by many governments about what do we want the norms of cyber to look like given the emergence of this capability.","startTime":2364.88,"endTime":2381.839,"durationSeconds":17,"level":"C1","overallScore":7.8,"rationale":"정책 논점과 구조 표현이 함께 많음."},{"segmentIndex":8,"text":"It's arriving at some kind of regulatory format which actually controls the industry in a way that helps people feel like it has democratic legitimacy rather than just a bunch of um extremely big personalities making big personality idiosyncratic decisions about perhaps the most powerful technology that's ever been built.","startTime":2463.28,"endTime":2483.28,"durationSeconds":20,"level":"C1","overallScore":7.8,"rationale":"민주적 정당성의 규제 틀을 제안함."},{"segmentIndex":11,"text":"very basic level is transparency force there to be labeling on AI companies and AI systems just as there is on food on children's toys on everything else safety testing some form of testing where third parties test out your technology for safety before it goes out into the world and then you're going to need to arrive at some scheme whereby if AI deployers have some number of safety incidents which is above some level or above some sort level of harm which is caused, there should be consequences for that.","startTime":2490.8,"endTime":2523.44,"durationSeconds":33,"level":"C1","overallScore":7.8,"rationale":"규제 방법을 구체적으로 제시함."},{"segmentIndex":33,"text":"to translate that to the classroom, you're going to need to some world where you are encouraging people to work without AI on primary source material, you know, books, scientific um education, education books, any of these primary sources that we use to learn generate their understanding of the thing and then I think you can have them talk with the AI and it suddenly scales the ability of that individual teacher to help all of these people because rather than having to pick among the 20 students in their classroom, each student gets personalized teaching, but the actual job of a teacher is to keep the AI out of the classroom for like a good hour while people read the primary source text and come up with their explanation and then let it back in.","startTime":2668,"endTime":2716.24,"durationSeconds":48,"level":"C1","overallScore":7.8,"rationale":"교실 활용 원칙을 구체적으로 설명함."},{"segmentIndex":65,"text":"There are these amazing dividends that have like come to the world through technology and in many places have started to work directly on some of our hardest problems ranging from climate to conservation.","startTime":3584.559,"endTime":3595.04,"durationSeconds":10,"level":"C1","overallScore":7.4,"rationale":"기술의 사회적 효과를 폭넓게 정리함."},{"segmentIndex":4,"text":"um we are dealing with the most powerful technology that's ever been built that is uh regulated less than any powerful technology before it and that situation is not like a long-term stable thing.","startTime":3618.24,"endTime":3628.88,"durationSeconds":11,"level":"C1","overallScore":7.6,"rationale":"위험성과 규제 문제를 구조적으로 설명."},{"segmentIndex":3,"text":"If there's anything that's going to need to change in the future, it's arriving at some kind of regulatory format which actually controls the industry rather than just a bunch of extremely big personalities making big personality idiosyncratic decisions about perhaps the most powerful technology that's ever been built.","startTime":23.6,"endTime":41.52,"durationSeconds":18,"level":"C2","overallScore":7,"rationale":"규제 필요성을 길게 논하는 통찰형 문장."},{"segmentIndex":10,"text":"I want to give the world a better way to understand the most powerful technology being developed in Silicon Valley and new ways to govern the private sector actors which are building that technology.","startTime":94.15899999999999,"endTime":105.119,"durationSeconds":11,"level":"C1","overallScore":7,"rationale":"목표와 규제 방향을 명확히 제시함."},{"segmentIndex":15,"text":">> Everyone who's building powerful technology throughout history whether electricity or airplanes or people who were building the first steam engines they had some intuitions about how it would change society and how it would lead to positive things but how it also had risk contained within it.","startTime":155.36,"endTime":171.76,"durationSeconds":16,"level":"C1","overallScore":6.8,"rationale":"역사적 일반화를 통해 교훈을 제시함."},{"segmentIndex":28,"text":"There are always uh unknown unknowns, but you can give yourself a much better ability to predict the future by trying really hard to imagine bad things that could happen and build tests for them ahead of time.","startTime":275.12,"endTime":289.04,"durationSeconds":14,"level":"C1","overallScore":7,"rationale":"미리 상상하고 시험하라는 핵심 교훈."},{"segmentIndex":34,"text":"But when Mythos came along, we pre-built all of these tests and immediately after the model arrived, we ran it through them and that instantly showed us that the AI system had got dramatically better at one class of risk.","startTime":324.88,"endTime":338.8,"durationSeconds":14,"level":"C1","overallScore":6.6,"rationale":"사전 테스트로 위험 증가를 확인한 사례."},{"segmentIndex":37,"text":"you know that I'm quite surprised given the advances in AI uh seem quite obvious but a lot of people are saying well this is all being overblown and it's obviously in their commercial interest to make mythos sound absolutely amazing because then everybody will want to buy it >> the unfortunate thing about AI is that AI systems get much better much more quickly than our intuitions do about how capable they are and I've been involved in this for many years so you know back in 2019 19, Dario, the CEO of Anthropic and myself, and we previously worked at OpenAI, rolled out the release of an early text generator called GPT2, which is a precursor to chat GPT.","startTime":354.32,"endTime":395.039,"durationSeconds":41,"level":"C1","overallScore":6.8,"rationale":"AI 발전 속도와 인식 차이를 설명함."},{"segmentIndex":57,"text":">> This feels similar to the runup to Y2K where you have this potential bug, right, where software hasn't been built to roll over from 1999 to 2000 and you don't know what will happen.","startTime":520,"endTime":532.72,"durationSeconds":13,"level":"C1","overallScore":7,"rationale":"Y2K 비유로 불확실성을 설명함."}],"generatedAt":"2026-06-26T16:41:26.490Z","keyClipsTotalSec":1457}]}