{"database": "press", "table": "releases", "rows": [["https://www.murphy.senate.gov/newsroom/press-releases/murphy-deems-artificial-intelligence-an-existential-threat-rejects-republican-fearmongering-calls-for-immediate-action-heading-into-the-midterms", "Murphy Deems Artificial Intelligence an 'Existential Threat', Rejects Republican Fearmongering, Calls for Immediate Action Heading Into the Midterms", "2026-10-27", "2026", "2026-10", "Democrat", "Senate", "CT", "Christopher Murphy", "M001169", "www.murphy.senate.gov", "chrismurphy", "https://www.murphy.senate.gov/newsroom/press-releases", "scraper", "WASHINGTON, D.C.\u2014Ahead of next week\u2019s midterm elections, U.S. Senator Chris Murphy (D-Conn.) sat down for a long-form interview with NBC News\u2019s Garrett Haake on the politics of artificial intelligence (AI) and the path ahead for Congress to regulate it. In a wide-ranging conversation, Murphy discussed the corrupt influence Big Tech currently holds over AI and data center policy, the urgent need to regulate AI before humans lose control over the technology, and his own bipartisan proposal, the AI Agent Accountability Act, which would force AI operators and developers to prioritize safety or face consequences including potential jail time when their technology goes rogue.\n\nSenator Murphy has been a leading voice in the U.S. Senate and on the global stage on the threat unregulated artificial intelligence poses to the cultural and economic health of the United States. In 2025, Murphy was named to the TIME AI 100 list as one of the most influential people shaping the conversation on AI.\n\nMurphy vowed that Democrats will make regulating the AI industry a top priority in the coming Congress: \u201cWe did a lot when Democrats were in charge, and it was a mistake for us not to add to a very long list of things we did, regulation of AI. If Democrats win control of the House and the Senate, then we will [regulate AI]. I mean, I can almost guarantee you we will pass legislation, at least beginning the work of regulating artificial intelligence.\u201d\n\nMurphy shut down the AI industry\u2019s claim that Congress doesn\u2019t have the technical expertise to regulate artificial intelligence: \u201cI don't buy that argument at all. Frankly, some of my colleagues make that argument: \u2018Oh, well, Congress isn't equipped to regulate AI.\u2019 I also don't know how to create nuclear energy. I don't know how to perform a surgery. But I regulate the health care industry. I am in charge of making sure that we're safe from nuclear disaster. I mean, if members of Congress weren't allowed to regulate industries that they aren't experts in, or industries in which they don't use the products, we wouldn't do anything. That's what a democracy is: you have citizens standing in, learning as much as you can, relying on experts, and making the best decisions that we can.\u201d\n\nMurphy emphasized there is ample precedent for regulating complex technologies that pose an existential threat to humanity: \u201cIt is true that we have had other potentially life-destroying inventions in the last 100 years: nuclear energy, for example, advanced medicines. And very quickly we decided that we were going to regulate those new technologies. We were going to accept the benefit of nuclear energy, but we were going to protect ourselves from the potential existential downside. On advanced medicines, we said, boy, if we don't regulate these things, they could kill us at scale. So let's get the good side of medicine and drug development and protect against the downside. That's possible with AI. There's this doomsday conversation, but only if we don't regulate it.\u201d\n\nMurphy pointed to several easy, common sense AI rules that Congress should pass immediately: \u201cI mean, some of it doesn't require you to know anything about technology. For instance, we shouldn't have deepfakes in our politics. We shouldn't have uncertainty every single minute of every single day as to whether what you're seeing is robot-created or is actual human content. We could just make that decision today, and you don't need any expertise to decide that we should ban deepfakes or watermark all AI-created content.\u201d\n\nMurphy debunked the \u2018China race\u2019 rhetoric coming from the Trump administration: \u201cWhat does that even mean? It is an industry-created red herring to prevent Congress from regulating the companies so that they can make trillions and trillions of dollars in profit. Why do we know that? Because China is regulating AI. China does not submit its citizens as guinea pigs to the technology. China right now, for all its flaws, is much more careful than the U.S. is, because they worry that if they unleash AI at scale on their economy, they will have hundreds of millions of job losses and immense political instability that comes with it. China is actually rooting for us to not regulate AI. So the companies say, no, China is rooting for us to regulate AI. It's exactly the opposite. China would love for us to disintegrate economically and politically before they do, because we choose not to protect people from artificial intelligence.\u201d\n\nMurphy also addressed the terrifying revelations from whistleblowers at Anthropic: \u201cIt could be that the most immediate problem here is these models that are escaping containment. We haven't really talked about that yet, but right now you are on the verge of having AI company-built models that escape human control. And they can't seem to fix that on their own. Why? Because they have to just speed so fast with research, development, and commercialization because of pressure from their investors that they basically need government, and they're kind of begging for government, to come in and say, you're going to have to start permitting our models. So that is a place where I think we could get some. Again, I think this only happens if Democrats are in charge. But if Democrats are in charge, there's a possibility that we could get legislation that controls these models, that approves these models before they get too far outside of the workshop.\u201d\n\nAs AI hacks of banks, governments, and businesses dominate the headlines, Murphy discussed his own bill to regulate the development and deployment of the AI agents at the center of these hacks: \u201cWe just introduced two weeks ago a bill that would create clear liability for the companies if their AI model escaped containment and did damage in any way, shape, or form. I think that's probably the law today, but it's really unclear because we don't have a body of case law. So that analogy is apt, but only if you have clear liability, and that's something that maybe we can get done at the beginning of next year as well.\u201d\n\nMurphy stressed the existential threat of AI goes much deeper than data centers or job loss: \u201cWhat happens to us as human beings if we start outsourcing our basic functionalities as humans, like creativity and friendship and conversation and problem solving, to machines? Kids report that they rely on AI for friendship and advice on a daily basis. And I'm not sure what's left of the human soul if machines start doing all those things for us. We become, very quickly, a very spiritually empty, unhappy, and anxious nation. And that will play out in political, cultural, social chaos as well. So it really is an existential threat. But again, only if we choose to let it go without any hindrance or check.\u201d\n\nMurphy\u2019s full interview is available here. A full transcript of Murphy\u2019s interview is available below.\n\nGarrett Haake: Welcome back to Here's the Scoop: Election Edition. I'm Garrett Haake. Connecticut's Democratic Senator Chris Murphy has become one of the most outspoken members of Congress when it comes to regulating AI. He recently sponsored a bipartisan bill with Republican Senator Josh Hawley to create civil and criminal liability for AI companies if their AI agents go rogue. As a member of the Foreign Relations Committee, he has a rare perspective on the AI race with China and how worried or not we should be about it. And as a dad to teenagers, he's voiced concerns about the societal and privacy impact of AI, especially of so-called AI friends. And as an in-demand campaign surrogate for other Democratic candidates this fall, he's been to the front lines of the debate over data centers sweeping the nation. All that means we have a lot to discuss. Chris Murphy, thanks for being here.\n\nSen. Chris Murphy: Yeah, awesome to be with you. Thanks for having me.\n\nGarrett Haake: I think a lot of people were kind of jolted into this AI debate back in September when we saw that post from a former Anthropic employee basically saying that they were concerned that, if left unattended, this technology could, like, destroy the world. Other AI companies jumped in and said, yeah, we agree, and they really weren't being regulated. You posted a video about it on X and you wrote, \u201cThe bottom line is we are cooked as an economy and a culture if we leave it to the industry to protect us. They won't, and they can't.\u201d Explain why you feel that way and who should be regulating AI.\n\nSen. Chris Murphy: So I think that AI in particular, this coming generation of AI that is going to essentially, you know, be super genius-level intelligence, is fundamentally different than any other technology that man has ever created. It's not just another version of the printing press. It is something brand new, and it's an existential threat.\n\nBut it is true that we have had other potentially life-destroying inventions in the last 100 years: nuclear energy, for example, advanced medicines. And very quickly we decided that we were going to regulate those new technologies. We were going to accept the benefit of nuclear energy, but we were going to protect ourselves from the potential existential downside. On advanced medicines, we said, boy, if we don't regulate these things, they could kill us at scale. So let's get the good side of medicine and drug development and protect against the downside. That's possible with AI. There's this doomsday conversation, but only if we don't regulate it.\n\nGarrett Haake: I mean, I covered the Hill for a long time. I've been in all these hearings about social media, other technological issues. Some of your colleagues in the Senate and the House don't send their own emails or post to their own social media accounts. Is Congress up to this challenge?\n\nSen. Chris Murphy: I don't buy that argument at all. Frankly, some of my colleagues make that argument: \u201cOh, well, Congress isn't equipped to regulate AI.\u201d I also don't know how to create nuclear energy. I don't know how to perform a surgery. But I regulate the health care industry. I am in charge of making sure that we're safe from nuclear disaster. I mean, if members of Congress weren't allowed to regulate industries that they aren't experts in, or industries in which they don't use the products, we wouldn't do anything. That's what a democracy is: you have citizens standing in, learning as much as you can, relying on experts, and making the best decisions that we can. So I just think that's bull****. Like, yeah, there's a lot of my colleagues that are old, and they don't use social media. But that doesn't mean that we're not smart enough to figure out, with the help of experts, how to properly protect us from this technology.\n\nI mean, some of it doesn't require you to know anything about technology. For instance, we shouldn't have deepfakes in our politics. We shouldn't have uncertainty every single minute of every single day as to whether what you're seeing is robot-created or is actual human content. We could just make that decision today, and you don't need any expertise to decide that we should ban deepfakes or watermark all AI-created content.\n\nGarrett Haake: But to your point, if it's all a choice, something like banning deepfakes in politics seems like it would be a layup for Congress to take on, and you haven't.\n\nSen. Chris Murphy: Yeah, I think there's a lot of reasons for that. First of all, we just have broad gridlock when Republicans are in charge. I mean, we're not doing anything right now, because the Trump administration and the Republican Congress is just a daily disaster of organization. We did a lot when Democrats were in charge, and it was a mistake for us not to add to a very long list of things we did, regulation of AI. But if Democrats win control of the House and the Senate, then we will. I mean, I can almost guarantee you we will pass legislation, at least beginning the work of regulating artificial intelligence.\n\nNow, part of the problem is there's a lot of money in these elections, a lot of AI industry money, that is attempting to elect members of Congress who will do nothing. That's what Elon Musk's money is in these races to do: to try to elect members of Congress who won't regulate AI, so that Musk and others can make as much money without guardrails as possible.\n\nGarrett Haake: What do you see as kind of the actual potential threats of unregulated AI versus what's sort of the science fiction threats?\n\nSen. Chris Murphy: Yeah, I mean, I don't know that any of them are science fiction threats any longer. But clearly, for many people, the most acute threat is disintegration of jobs. And I think that's going to happen sooner rather than later. I think you will see the wholesale elimination of certain categories of jobs in the next five to 10 years. And as a parent of a kid who's going to college, you know, I know the conversation that parents are having today: What is my kid going to do? What should I send them to school to learn?\n\nSecond is the disintegration of our politics. You're already seeing that with a lot of campaigns running these deepfake ads. It makes it a lot easier for outside players, foreign governments, to be able to send fake information into our political ecosystem and just tear us apart.\n\nBut I think the most significant downside is the spiritual impact of AI, because what happens to us as human beings if we start outsourcing our basic functionalities as humans, like creativity and friendship and conversation and problem solving, to machines? Kids report that they rely on AI for friendship and advice on a daily basis. And I'm not sure what's left of the human soul if machines start doing all those things for us. We become, very quickly, a very spiritually empty, unhappy, and anxious nation. And that will play out in political, cultural, social chaos as well. So it really is an existential threat. But again, only if we choose to let it go without any hindrance or check.\n\nGarrett Haake: I think that sort of humanist perspective on it is interesting. It sort of creates strange political bedfellows, right? I mean, I mentioned you and Josh Hawley working together on a bill on rogue AI agents. I don't think of you and Josh Hawley as aligned on a traditional political spectrum on much of anything. But can you talk a little bit about how these AI-related issues have rearranged the political spectrum a little bit?\n\nSen. Chris Murphy: Yeah, I think if you listen to the conversation happening on the right, you will hear a lot of this worry about the spiritual cost of artificial intelligence replacing human functionality. Josh Hawley and I are working on that piece of legislation, but also a bill to effectively ban kids from having access to these chatbots. Maybe that's because we're both parents, but I think it's also because inside the Republican Party is a broad skepticism of technological elites. It kind of got buried once Trump came into office, because he decided to just celebrate the tech elites, the folks who are running these companies. But inside that party, you know, there's always been skepticism of the folks that are running the big information technology companies.\n\nGarrett Haake: What do you make of the way that the president has handled this issue? I mean, his political radar, whatever you think of his policies, is usually pretty good, but he seems to have found himself almost alone on the 20 side of, like, an 80/20 issue. I guess a 70/30 issue in our poll. Like, I mean, 70% of Americans are distrustful of AI, so they're more concerned about it than excited about it, and he's kind of over there acting as the industry's biggest cheerleader. What do you attribute that to?\n\nSen. Chris Murphy: Yes, he is very badly on the wrong side of this issue, and his embrace of data centers, you know, is probably going to be a really big issue once people show up to the polls. But he's blinded by his desire to hold power forever, and these companies, by funneling so much money into his political operations and supporting so many of his favorite political candidates, I think, are causing him to, you know, not care or not see the enormous political downside.\n\nGarrett Haake: I feel like the AI debate more broadly has been kind of subsumed of late with discussion just about data centers specifically. You've been traveling all over the country. You've been watching, I think, how candidates are talking about this. Who do you think is addressing the idea of data centers in communities particularly well?\n\nSen. Chris Murphy: I mean, Sherrod Brown, you know, has always been ahead of the curve, and Sherrod Brown was talking about the danger of these data centers and the danger of unregulated AI before anybody else was. You know, I saw Mayor Mamdani give a speech yesterday or this week in which he talked, like I talk, about the spiritual impact of AI. And I do think that it's important to speak to people both about the jobs impact and the impact on human dignity.\n\nThe other piece that really comes out in the way that people talk about data centers is the prism of corruption. You know, if you look, people really believe that these data centers are being located in their communities because of a corrupt, behind-the-scenes deal being done between the companies and their elected leaders. I think that is true in some of the cases, but the corruption angle is probably not as widespread in the location of data centers as people think. But what it's telling you is that people think that the whole conversation about AI is corrupt. They know that these companies are flooding the zone with money. They know these companies get to speak to members of Congress in a way that they don't. And it's important to understand that they believe that the whole political system is corrupt. And until you fix the political system, until you do something about the increased input that rich people have into our politics, they don't really believe that any regulatory system is going to be effective.\n\nGarrett Haake: Your own governor, Ned Lamont, is getting kind of beaten up for having pushed for tax breaks for some of these data centers in Connecticut. Now he seems to be walking that back a little bit. He's not called for a moratorium in Connecticut on their construction. Do you think he should?\n\nSen. Chris Murphy: Oh, listen, I don't know that you need to have a moratorium if you have a real certain sense of how these data centers are going to impact the community. But I also think you have to have a firm federal regulatory structure. In the absence of a federal regulatory structure, it may be that a moratorium makes sense, because it's your only way, for instance, to stop the wholesale jobs destruction. If AI companies have a harder time building data centers, then they just aren't going to be able, for instance, to put out of business millions of coders in this country.\n\nSo ultimately, I think what you have to have is a licensing system, both for models and for products, especially, or at least, in key sensitive industries, partnered with a tougher regulatory system around how you locate data centers. So a moratorium is probably not necessary as long as you have a federal regulatory system, but we don't have that yet. So for a lot of states and a lot of leaders and a lot of constituencies, though a moratorium is a blunt tool, it's probably your only way to stop the job destruction.\n\nGarrett Haake: I have two more things I want to make sure we talk about before we run out of time. The first is an argument that I hear all the time when it comes to regulation around AI: that somehow the U.S. cannot regulate or slow down these companies, or China will win the AI race.\n\nSen. Chris Murphy: What does that even mean? It is an industry-created red herring to prevent Congress from regulating the companies so that they can make trillions and trillions of dollars in profit. Why do we know that? Because China is regulating AI. China does not submit its citizens as guinea pigs to the technology. China right now, for all its flaws, is much more careful than the U.S. is, because they worry that if they unleash AI at scale on their economy, they will have hundreds of millions of job losses and immense political instability that comes with it. China is actually rooting for us to not regulate AI. So the companies say, no, China is rooting for us to regulate AI. It's exactly the opposite. China would love for us to disintegrate economically and politically before they do, because we choose not to protect people from artificial intelligence.\n\nSo just spend a couple of hours researching China's AI regulatory scheme, and you will see that they are light-years ahead of the United States, because they actually are hoping that we do nothing and that AI destroys us before it destroys them. And their data center construction is a fraction of ours, despite the fact that in China there is no such thing as, like, a NIMBY movement. I mean, the Chinese government could snap their fingers and say we're going to build a thousand of these things, and they haven't. Why not? Well, because China also is investing in models that don't use as much compute power, whereas the United States is totally agnostic. We basically have told the companies, you know, you can develop your models in any way, shape, or form. The way they're doing it involves massive compute power, which is why there's a real risk of our electricity price going up. China instead has decided to create models that need fewer data centers. That sounds like a really smart thing that, if China is deciding to do it, maybe the United States should decide that as well.\n\nGarrett Haake: Stop me if this is a terrible metaphor, but it makes me think of, like, the difference between the U.S. and the Japanese car companies, where the U.S. is sort of building, like, the biggest, loudest, most gas-guzzling version, and the Japanese like, what if we did this smaller and more efficient and sold them to your market? I mean, is that a fair metaphorical comparison?\n\nSen. Chris Murphy: It's probably a fair analogy. I think what we need in this country broadly is industrial policy. We are going to succeed because we are a free market economy, and I think ultimately China's command, state-run economy can't match ours. So we want to be careful in terms of how much direction we give the AI industry from government. But we certainly need more integration between the public sector and the private sector in order to compete with China.\n\nGarrett Haake: The other thing I want to make sure to ask you about is, we talked about deepfakes on the campaign trail, but how else are you seeing campaigns make use of AI? Because I could see that technology being transformative for all the different things that a good campaign has to do to get someone elected.\n\nSen. Chris Murphy: No, it certainly will supersize a lot of the targeting efforts that have already been underway for decades in American politics, where, you know, you're really looking hard for the selection of voters that are going to decide an election, either the people who need to be persuaded or the people that need to be turned out. And so the AI tools are certainly going to help you create content, but they're also going to help you microtarget your message. It's not just AI video content. There's going to be tons of other AI-created content on, you know, more text-based social media platforms, and more content that looks like a real human being, more content that interacts with users and other actual humans that are on those sites, like Twitter or Truth Social or Bluesky. So there's just going to be a lot more fakery. There's going to be a lot more sort of microtargeted messaging. I'm not sure that any of that is really good for our politics. But yeah, it's in the middle of transforming the political world right now.\n\nGarrett Haake: So when you guys come back in January, is there a piece of kind of low-hanging fruit that a potentially divided Congress could pass on this issue and that President Trump could sign? They would start this process of regulating AI.\n\nSen. Chris Murphy: So I've identified in this conversation a couple pieces of consumer- and family-facing low-hanging fruit that would be, you know, regulating or restricting these friendship bots, taking on the deepfakes. But it could be that the most immediate problem here is these models that are escaping containment. We haven't really talked about that yet, but right now you are on the verge of having AI company-built models that escape human control. And they can't seem to fix that on their own. Why? Because they have to just speed so fast with research, development, and commercialization because of pressure from their investors that they basically need government, and they're kind of begging for government, to come in and say, you're going to have to start permitting our models. So that is a place where I think we could get some. Again, I think this only happens if Democrats are in charge. But if Democrats are in charge, there's a possibility that we could get legislation that controls these models, that approves these models before they get too far outside of the workshop.\n\nGarrett Haake: The containment thing is wild to me, because I feel like if you were, like, a nation full of zoo operators whose lions kept getting out, governments would not tolerate that. But because it's invisible bots, we're like, oh yeah, it's fine.\n\nSen. Chris Murphy: Well, government wouldn't tolerate that, but the zoos wouldn't tolerate that either, because they know that they would lose the zoo. They'd get sued the minute that a lion ate a bunch of school kids. So that's what Josh Hawley's bill that he has with me does. We just introduced two weeks ago a bill that would create clear liability for the companies if their AI model escaped containment and did damage in any way, shape, or form. I think that's probably the law today, but it's really unclear because we don't have a body of case law. So that analogy is apt, but only if you have clear liability, and that's something that maybe we can get done at the beginning of next year as well.", 1, "2026-11-09T13:36:59Z", "2026-11-09T13:37:55Z"]], "columns": ["url", "title", "date", "year", "month", "party", "chamber", "state", "member_name", "bioguide_id", "domain", "scraper", "source", "date_source", "text", "has_text", "collected_at", "updated_at"], "primary_keys": ["url"], "primary_key_values": ["https://www.murphy.senate.gov/newsroom/press-releases/murphy-deems-artificial-intelligence-an-existential-threat-rejects-republican-fearmongering-calls-for-immediate-action-heading-into-the-midterms"], "units": {}, "query_ms": 0.7897249888628721, "source": "dwillis/congress-press", "source_url": "https://github.com/dwillis/congress-press", "license": "MIT", "license_url": "https://github.com/dwillis/congress-press/blob/main/LICENSE"}