Sept. 11, 2026
AI Headlines Just Got Scarier: Anthropic Disrupted Potential Plans To Build Biological Weapons!
Apple Podcasts 
Spotify 
Castro 
RSS Feed A new report from Anthropic is adding new concerns about the growing dangers of artificial intelligence, highlighting the real-life concerns after one if its lead researchers warned of AI’s potential to destroy all of mankind by the end of the decade. Anthropic says over the past eight months, it’s identified and disrupted dozens of potential plots to use its technology “for malicious activity.” That includes blocking scientists who have used its AI models to conduct research that could have developed biological weapons. Experts say the development of biological pathogens is among the gravest concerns they have when it comes to the rapidly growing technology.
See omnystudio.com/listener for privacy information.
00:00:02
Speaker 1: What is up with all of these AI headlines? I thought it was bad when we heard that basically all of mankind might be eviscerated by the end of the decade, but there are even scarier headlines out there today talking about biological weapons of warfare and the fact that Anthropic, yes, the AI company that's been at the center of all of these headlines with one of its former researchers making some of these doomsday predictions, they actually just released a report that acknowledges over the past eight months, they've disrupted dozens of operations where they have folks who they believe may have been trying to use their bot, their AI bot, for malicious activity. This is scary stuff, everyone, and welcome with that fun headline. Welcome to Amy and TJ. It's Friday, September 11th. It's just now it feels like, Pandora's box has been opened and we're finding out all sorts of stuff that maybe I didn't.
00:01:06
Speaker 2: Want to know.
00:01:07
Speaker 3: There's nowhere to hide. We can't even leave the country and go somewhere in safety. This is everywhere. There's nowhere to hide from this. They are actually talking about it. This sounds so far-fetched. It sounds like something out of science fiction because this is the only place we've seen this in movies. That some device, some self-learning, some AI, super intelligent something is out of human control starts to learn on its own, develop on its own, and we have no idea how or way to stop it. Yeah, that was the past couple of days. Now you're talking about bioweapons. This story, I'd be honest with you, I struggle to understand because this AI stuff, they think somebody did something, but they can't confirm somebody did something. And even if somebody did something, we can't prove if it was a nefarious intent or not. They don't even know.
00:01:57
Speaker 2: That's where we are.
00:01:58
Speaker 1: And look, yes, so the headlines before were that the AI, AI itself was going to make new versions of itself that we wouldn't be able to control that would basically kill everybody. But this report is showing that, yes, there are human beings behind some of this ill intent and that they're a threat intelligence team. The fact that they've got a whole team dedicated to try and find folks out there and they named countries like China. Iran that are out there actively trying to use AI technology to develop bio weapons to use as.
00:02:37
Speaker 4: Attacks against our country and whoever else.
00:02:40
Speaker 2: They don't like. And this is scary.
00:02:42
Speaker 1: They say that some of these scientists, they believe, are state-sponsored. So this is absolutely folks who they believe had malintent and were intent on using it for harm.
00:02:54
Speaker 3: Yeah, to pit this, so the first headlines this week, right, Robes? This happened pretty quickly. So they go from telling us that the superintelligence, the AI is developing at such a rapid rate that it could lead to the end of humanity. By the end of the decade, one guy comes out and says that, others back him up. Yes, this is our fear. So, Robes, the question was, well, how could this end the world? How could this end all human mankind? And they gave a couple of scenarios. One was that this super intelligence develops on its own, gets out of control, and starts to do what it wants to do and launches nuclear weapons, whatever. And the other option they gave us, Robes, bad actors could use AI to possibly start to develop Some biological nuclear weapons. And then a couple days later, they send out a report and say, actually... It's happened.
00:03:44
Speaker 1: Yeah.
00:03:44
Speaker 2: And it's been happening.
00:03:45
Speaker 4: That is.
00:03:46
Speaker 3: Okay, that's problematic.
00:03:47
Speaker 1: Yes, to hear that basically they said dozens of instances of potentially malicious use of its AI have taken place since December of 2025. Yes, in an eight-month period. And so they didn't disclose names of the scientists or... their countries of affiliation in the particular bioweapons threat. But they did say that they were deeply concerned. Like you said earlier, they can't prove that they were going to use it for weapons of mass destruction, basically, because they said, okay, technically, they could have been trying to use AI for a breakthrough in a vaccine or to fight some sort of an infectious disease. But they also said that it was very likely that these groups were actually trying to engineer AI or create a dangerous pathogen. To think about that is beyond scary, and so they claim they erred on the side of caution by blocking these scientists. But to see this in a report, that in the last eight months, there have been dozens of attempts to use AI to create and manufacture biological weapons, that is frightening.
00:05:01
Speaker 3: The more frightening, Robes, They can't, or excuse me, did you get the impression they actually know and didn't put it in the report of who these folks are?
00:05:10
Speaker 2: Yes, because they blocked them.
00:05:12
Speaker 1: They know who they are, and they blocked them from using their information.
00:05:16
Speaker 2: AI tool.
00:05:17
Speaker 3: But they know these folks by name and identity, not just it's coming from this direction and we don't know who's other.
00:05:22
Speaker 2: They know who they are.
00:05:23
Speaker 1: They say they chose not to disclose the names of the researchers or the institutions they were working in.
00:05:30
Speaker 3: Because?
00:05:31
Speaker 1: Because they have uncertainty about what their motives were. Yes.
00:05:35
Speaker 3: That's crazy to me.
00:05:36
Speaker 2: They don't know their motives. That is concerning.
00:05:40
Speaker 3: So we don't know if these folks actually were up to something is what they're saying.
00:05:45
Speaker 2: That's exactly what they're saying.
00:05:46
Speaker 1: But it seems as though the way it was worded, I read much of the report, they're really suspicious that they were. Like they believe that they almost certainly were, but they can't necessarily prove it.
00:05:58
Speaker 2: That was the way it was worded.
00:06:00
Speaker 1: And look, they even said that the potential, like we've talked about all the different scenarios about what AI could do. They actually say the idea of some foreign agent creating a biological pathogen is among the most concerning possibilities that experts have because it's among the most likely.
00:06:19
Speaker 2: It's among the most attainable and quickly attainable.
00:06:22
Speaker 1: So that was what was disturbing because we were hearing about all these threats and for them to say, actually, this biological warfare thing is actually one of the most scary of them all because it's the most likely. And look, they went into some of the cases where they know there was malintent They say it ranged from a network of fake dating apps designed to defraud users to surveillance systems that were built to identify and monitor dissidents. So they're talking about Iran, like you're trying to find folks who are rallying against your government or folks who might be protesting against policies.
00:06:57
Speaker 2: Well, they can use AI to surveil and monitor.
00:07:00
Speaker 4: Folks, which is completely illegal, to then take them out basically.
00:07:04
Speaker 1: So using AI, obviously, for bad, tyrannical reasons versus to create.
00:07:14
Speaker 2: New vaccines.
00:07:15
Speaker 1: There's so much good that can come from this, so many medical breakthroughs that could result from AI, but unfortunately, with the good comes the absolutely evil, horrible, and perhaps the end of humanity.
00:07:29
Speaker 3: Robes, as we were sitting here talking, New York Times alert, just email came across their new opinion piece. When will we finally regulate AI? Conversations are going to change. It does. It sounds like science fiction until you read past the headline. And it'll scare the hell out of you.
00:07:49
Speaker 2: It does.
00:07:50
Speaker 1: It should.
00:07:51
Speaker 3: We got the smartest people working on the most advanced technology. The stuff they're creating, they're coming out and saying they're scared of.
00:07:59
Speaker 4: Yes.
00:08:00
Speaker 1: I mean, I don't think it gets scarier than that. And look, we're talking about the big fear of biological weapons, of nuclear warfare, all of that. But they also talked about, and these are known highlighted cases that are in this report that was just released by Anthropic, saying they can confirm Russian state media using CLAWD to generate propaganda that masquerades as independent reporting. So yes, all of this, you can't believe what you read necessarily. Sourcing has never become more important. Anytime you see sources say, and even if there's like nebulous sources, you can't trust it because AI is behind a lot of it. And it's being used specifically by countries who want to put out misinformation to either sway folks into their favor or just to alarm folks in different directions. They say there were attempts in China, Russia, and Yemen to use Claude to develop software for conventional weapons development, including firearms, missiles, armed drones, and bombs. I mean, this is beyond frightening, and that is absolutely verifiable.
00:09:07
Speaker 3: In the last eight months, babe, I guess we're going to start paying much more attention to this. This is one, Robes, I don't get that. This is one where I am quite active in wanting to know what Congress is doing about this. Because this is completely out of control technology. Robes, we talk in the Lindsay Clancy case previously, like the battle of the experts, which one do you believe? Robes, if all of them got up there and said the same thing... Which is what they're doing right now. This is what all of these researchers and AI are doing. Saying the exact same scary as hell thing. Robes, Anthropic didn't even come out and challenge what this guy who just resigned. They didn't even. Robes, they are scaring the hell out of me because I've watched Terminator 1, 2, 3, 4, 5. Yes. Robes, Skynet takes over. August 29th, billions die. Robes, this is exactly the scenario.
00:10:12
Speaker 4: And you know what, babe?
00:10:13
Speaker 2: Fear is motivating.
00:10:15
Speaker 1: And maybe Lindsay Clancy tore us apart. Maybe just about everything tears us apart. But would you believe that this new AI threat is bringing together some of the most unlikely political enemies? Welcome back, everyone, to Amy and TJ. Would you believe that this latest threat and these headlines that are scaring the heck out of us have actually brought together Ted Cruz.
00:10:48
Speaker 2: And Bernie Sanders?
00:10:51
Speaker 3: You know, that doesn't even sound right.
00:10:52
Speaker 4: I mean, I laughed saying it out loud.
00:10:54
Speaker 3: That doesn't sound right.
00:10:54
Speaker 1: But they have both been working on legislation that would bring some safety measures, some guardrails to artificial intelligence. You mentioned what folks have been saying.
00:11:05
Speaker 2: I don't recall this, but Ted Cruz.
00:11:09
Speaker 1: He was on The View, and he was talking about AI and the risk that it carries with it.
00:11:15
Speaker 4: And he said... that he had Elon.
00:11:18
Speaker 1: Musk on his podcast last year, and he recalled what Elon Musk told him last year. I don't know how this didn't make headlines. I don't know why we weren't thinking about this last year. He said he asked him about AI, and he said, what are the odds that AI destroys humanity? He said he thought he was just kind of throwing out a funny question. He said that Elon Musk told him and said 10% to 20%. And that is exactly the percentages we're hearing from researchers this week who are actively a part of this artificial intelligence boom.
00:11:54
Speaker 3: You said it right. He thought he was throwing out a silly question. It sounds like a throwaway. It sounds so fantastical that, no, of course nothing like that can happen. They're telling us it's happening.
00:12:07
Speaker 1: AI.
00:12:08
Speaker 3: has been doing things on its own in some very high profile cases. You tell a machine to do one thing and it goes and does something else on its own that you didn't tell it to and you didn't know it did. It hacked into other companies that it wasn't asked to do. Robes, that is the proof right there. What if it would have hacked into, I don't know, the energy grid?
00:12:29
Speaker 2: Or our banking system.
00:12:31
Speaker 4: Yes.
00:12:32
Speaker 2: The stock market.
00:12:33
Speaker 3: It's out of control.
00:12:34
Speaker 2: It's out of control.
00:12:35
Speaker 1: And so, yes, we have to worry about AI creating new, scary superpower AIs. And we also have to worry about Iran and China and Russia and all of our sworn enemies also racing to use AI to do its bidding, to become its own powerful entity in and of itself. So It's like you're battling so many different real and unseen demons. You don't even know who to possibly try and guard against. And how do we police this? And that's the question. Lawmakers, yes, they want to create legislation. But the problem is, okay, so we slow down here in America. You can't stop China. You can't stop Russia. You can't stop any other country from actively pursuing their own AI.
00:13:21
Speaker 3: We've got to win the race. We've got to win the race to superintelligence. If we don't, somebody else will. And then we're in trouble. That's what we're doing.
00:13:29
Speaker 2: But we can't trust the superintelligence.
00:13:31
Speaker 1: So what we're racing to get to is actually probably the most untrustworthy of them all.
00:13:36
Speaker 3: As long as we get there first. That's all. And that's the problem. They say we're focused so much on that, that some safety or, yes, scenarios about the world ending are not being taken as seriously or given the priority that they should be.
00:13:50
Speaker 1: And look, you're going to be hearing about this now pretty much daily because.
00:13:53
Speaker 3: If you think- From us.
00:13:55
Speaker 4: From us, no, but also think about midterm elections.
00:13:57
Speaker 1: This is now- on the forefront of so many voters' minds, and they want to know what Congress is going to do. And not many of us have much hope that they'll do anything about it. But I was heartened to see that Bernie Sanders and Ted Cruz are on the same side of this.
00:14:14
Speaker 4: And why wouldn't they be?
00:14:16
Speaker 1: This should be, if you can think of one thing that would get us all in line on the same page with the same motive, it would be us against the world, us against super intelligence. If we can't come together on this, I don't know what we can come together on.
00:14:33
Speaker 3: We see it on movies all the time, right? It's the only time the whole world community comes together when there's a threat to human existence. This is one I'm being told. I mean, Armageddon, Independence Day. Aliens, yeah. The whole world comes together because we have a threat that's going to take out the entire earth, the human population. That is what we are talking about, Robes, And, folks, obviously we're not being alarmist. We are not scientists in there. But, Robes, we need to listen. These are the right people speaking that we have to talk to. And you can't be a layperson, Robes. Why can't you just go down there and pull the plug, right? Just kick it out of the socket and we can turn the power off to the AI. It's not how it works. Because the thing has the ability to duplicate outside of what other place. So now it can just get out. You can't just unplug a machine is what we're dealing with now. Because that was my first question. Why don't you just turn the power off?
00:15:31
Speaker 1: Oh my God. They might turn the power off on us first. No, but I just, I feel like you have that threat. And then just to know that Anthropic, one company, caught dozens of bad actors who Did they catch everybody?
00:15:47
Speaker 2: Did someone slip through the cracks?
00:15:48
Speaker 1: I mean, this is deeply concerning, and I know that there will be more reports and certainly more investigations, but the question is, can we actually do anything about this?
00:15:57
Speaker 2: I don't know how confident I feel.
00:15:59
Speaker 3: Enjoy your day. Enjoy your life. Enjoy your loved ones. There are just so many threats coming from all over the place. There's financial strain. There's emotional strain. There's just strain from society and not getting along with your neighbor. Now you're telling us the world might actually frickin' end because of computers. You all are desperate. to develop faster than China. Okay. How much fighting, how much more can we do here, babe? It's just exhausting.
00:16:23
Speaker 2: It is.
00:16:23
Speaker 4: It is indeed.
00:16:25
Speaker 1: We'll continue to follow the quickly evolving headlines coming out of artificial intelligence these days.
00:16:32
Speaker 2: But in the meantime, yes, enjoy your weekend. Enjoy your day. We always appreciate you listening to us.
00:16:36
Speaker 4: I'm Amy Robach alongside TJ Holmes.
00:16:39
Speaker 2: We'll talk soon.
Speaker 1: What is up with all of these AI headlines? I thought it was bad when we heard that basically all of mankind might be eviscerated by the end of the decade, but there are even scarier headlines out there today talking about biological weapons of warfare and the fact that Anthropic, yes, the AI company that's been at the center of all of these headlines with one of its former researchers making some of these doomsday predictions, they actually just released a report that acknowledges over the past eight months, they've disrupted dozens of operations where they have folks who they believe may have been trying to use their bot, their AI bot, for malicious activity. This is scary stuff, everyone, and welcome with that fun headline. Welcome to Amy and TJ. It's Friday, September 11th. It's just now it feels like, Pandora's box has been opened and we're finding out all sorts of stuff that maybe I didn't.
00:01:06
Speaker 2: Want to know.
00:01:07
Speaker 3: There's nowhere to hide. We can't even leave the country and go somewhere in safety. This is everywhere. There's nowhere to hide from this. They are actually talking about it. This sounds so far-fetched. It sounds like something out of science fiction because this is the only place we've seen this in movies. That some device, some self-learning, some AI, super intelligent something is out of human control starts to learn on its own, develop on its own, and we have no idea how or way to stop it. Yeah, that was the past couple of days. Now you're talking about bioweapons. This story, I'd be honest with you, I struggle to understand because this AI stuff, they think somebody did something, but they can't confirm somebody did something. And even if somebody did something, we can't prove if it was a nefarious intent or not. They don't even know.
00:01:57
Speaker 2: That's where we are.
00:01:58
Speaker 1: And look, yes, so the headlines before were that the AI, AI itself was going to make new versions of itself that we wouldn't be able to control that would basically kill everybody. But this report is showing that, yes, there are human beings behind some of this ill intent and that they're a threat intelligence team. The fact that they've got a whole team dedicated to try and find folks out there and they named countries like China. Iran that are out there actively trying to use AI technology to develop bio weapons to use as.
00:02:37
Speaker 4: Attacks against our country and whoever else.
00:02:40
Speaker 2: They don't like. And this is scary.
00:02:42
Speaker 1: They say that some of these scientists, they believe, are state-sponsored. So this is absolutely folks who they believe had malintent and were intent on using it for harm.
00:02:54
Speaker 3: Yeah, to pit this, so the first headlines this week, right, Robes? This happened pretty quickly. So they go from telling us that the superintelligence, the AI is developing at such a rapid rate that it could lead to the end of humanity. By the end of the decade, one guy comes out and says that, others back him up. Yes, this is our fear. So, Robes, the question was, well, how could this end the world? How could this end all human mankind? And they gave a couple of scenarios. One was that this super intelligence develops on its own, gets out of control, and starts to do what it wants to do and launches nuclear weapons, whatever. And the other option they gave us, Robes, bad actors could use AI to possibly start to develop Some biological nuclear weapons. And then a couple days later, they send out a report and say, actually... It's happened.
00:03:44
Speaker 1: Yeah.
00:03:44
Speaker 2: And it's been happening.
00:03:45
Speaker 4: That is.
00:03:46
Speaker 3: Okay, that's problematic.
00:03:47
Speaker 1: Yes, to hear that basically they said dozens of instances of potentially malicious use of its AI have taken place since December of 2025. Yes, in an eight-month period. And so they didn't disclose names of the scientists or... their countries of affiliation in the particular bioweapons threat. But they did say that they were deeply concerned. Like you said earlier, they can't prove that they were going to use it for weapons of mass destruction, basically, because they said, okay, technically, they could have been trying to use AI for a breakthrough in a vaccine or to fight some sort of an infectious disease. But they also said that it was very likely that these groups were actually trying to engineer AI or create a dangerous pathogen. To think about that is beyond scary, and so they claim they erred on the side of caution by blocking these scientists. But to see this in a report, that in the last eight months, there have been dozens of attempts to use AI to create and manufacture biological weapons, that is frightening.
00:05:01
Speaker 3: The more frightening, Robes, They can't, or excuse me, did you get the impression they actually know and didn't put it in the report of who these folks are?
00:05:10
Speaker 2: Yes, because they blocked them.
00:05:12
Speaker 1: They know who they are, and they blocked them from using their information.
00:05:16
Speaker 2: AI tool.
00:05:17
Speaker 3: But they know these folks by name and identity, not just it's coming from this direction and we don't know who's other.
00:05:22
Speaker 2: They know who they are.
00:05:23
Speaker 1: They say they chose not to disclose the names of the researchers or the institutions they were working in.
00:05:30
Speaker 3: Because?
00:05:31
Speaker 1: Because they have uncertainty about what their motives were. Yes.
00:05:35
Speaker 3: That's crazy to me.
00:05:36
Speaker 2: They don't know their motives. That is concerning.
00:05:40
Speaker 3: So we don't know if these folks actually were up to something is what they're saying.
00:05:45
Speaker 2: That's exactly what they're saying.
00:05:46
Speaker 1: But it seems as though the way it was worded, I read much of the report, they're really suspicious that they were. Like they believe that they almost certainly were, but they can't necessarily prove it.
00:05:58
Speaker 2: That was the way it was worded.
00:06:00
Speaker 1: And look, they even said that the potential, like we've talked about all the different scenarios about what AI could do. They actually say the idea of some foreign agent creating a biological pathogen is among the most concerning possibilities that experts have because it's among the most likely.
00:06:19
Speaker 2: It's among the most attainable and quickly attainable.
00:06:22
Speaker 1: So that was what was disturbing because we were hearing about all these threats and for them to say, actually, this biological warfare thing is actually one of the most scary of them all because it's the most likely. And look, they went into some of the cases where they know there was malintent They say it ranged from a network of fake dating apps designed to defraud users to surveillance systems that were built to identify and monitor dissidents. So they're talking about Iran, like you're trying to find folks who are rallying against your government or folks who might be protesting against policies.
00:06:57
Speaker 2: Well, they can use AI to surveil and monitor.
00:07:00
Speaker 4: Folks, which is completely illegal, to then take them out basically.
00:07:04
Speaker 1: So using AI, obviously, for bad, tyrannical reasons versus to create.
00:07:14
Speaker 2: New vaccines.
00:07:15
Speaker 1: There's so much good that can come from this, so many medical breakthroughs that could result from AI, but unfortunately, with the good comes the absolutely evil, horrible, and perhaps the end of humanity.
00:07:29
Speaker 3: Robes, as we were sitting here talking, New York Times alert, just email came across their new opinion piece. When will we finally regulate AI? Conversations are going to change. It does. It sounds like science fiction until you read past the headline. And it'll scare the hell out of you.
00:07:49
Speaker 2: It does.
00:07:50
Speaker 1: It should.
00:07:51
Speaker 3: We got the smartest people working on the most advanced technology. The stuff they're creating, they're coming out and saying they're scared of.
00:07:59
Speaker 4: Yes.
00:08:00
Speaker 1: I mean, I don't think it gets scarier than that. And look, we're talking about the big fear of biological weapons, of nuclear warfare, all of that. But they also talked about, and these are known highlighted cases that are in this report that was just released by Anthropic, saying they can confirm Russian state media using CLAWD to generate propaganda that masquerades as independent reporting. So yes, all of this, you can't believe what you read necessarily. Sourcing has never become more important. Anytime you see sources say, and even if there's like nebulous sources, you can't trust it because AI is behind a lot of it. And it's being used specifically by countries who want to put out misinformation to either sway folks into their favor or just to alarm folks in different directions. They say there were attempts in China, Russia, and Yemen to use Claude to develop software for conventional weapons development, including firearms, missiles, armed drones, and bombs. I mean, this is beyond frightening, and that is absolutely verifiable.
00:09:07
Speaker 3: In the last eight months, babe, I guess we're going to start paying much more attention to this. This is one, Robes, I don't get that. This is one where I am quite active in wanting to know what Congress is doing about this. Because this is completely out of control technology. Robes, we talk in the Lindsay Clancy case previously, like the battle of the experts, which one do you believe? Robes, if all of them got up there and said the same thing... Which is what they're doing right now. This is what all of these researchers and AI are doing. Saying the exact same scary as hell thing. Robes, Anthropic didn't even come out and challenge what this guy who just resigned. They didn't even. Robes, they are scaring the hell out of me because I've watched Terminator 1, 2, 3, 4, 5. Yes. Robes, Skynet takes over. August 29th, billions die. Robes, this is exactly the scenario.
00:10:12
Speaker 4: And you know what, babe?
00:10:13
Speaker 2: Fear is motivating.
00:10:15
Speaker 1: And maybe Lindsay Clancy tore us apart. Maybe just about everything tears us apart. But would you believe that this new AI threat is bringing together some of the most unlikely political enemies? Welcome back, everyone, to Amy and TJ. Would you believe that this latest threat and these headlines that are scaring the heck out of us have actually brought together Ted Cruz.
00:10:48
Speaker 2: And Bernie Sanders?
00:10:51
Speaker 3: You know, that doesn't even sound right.
00:10:52
Speaker 4: I mean, I laughed saying it out loud.
00:10:54
Speaker 3: That doesn't sound right.
00:10:54
Speaker 1: But they have both been working on legislation that would bring some safety measures, some guardrails to artificial intelligence. You mentioned what folks have been saying.
00:11:05
Speaker 2: I don't recall this, but Ted Cruz.
00:11:09
Speaker 1: He was on The View, and he was talking about AI and the risk that it carries with it.
00:11:15
Speaker 4: And he said... that he had Elon.
00:11:18
Speaker 1: Musk on his podcast last year, and he recalled what Elon Musk told him last year. I don't know how this didn't make headlines. I don't know why we weren't thinking about this last year. He said he asked him about AI, and he said, what are the odds that AI destroys humanity? He said he thought he was just kind of throwing out a funny question. He said that Elon Musk told him and said 10% to 20%. And that is exactly the percentages we're hearing from researchers this week who are actively a part of this artificial intelligence boom.
00:11:54
Speaker 3: You said it right. He thought he was throwing out a silly question. It sounds like a throwaway. It sounds so fantastical that, no, of course nothing like that can happen. They're telling us it's happening.
00:12:07
Speaker 1: AI.
00:12:08
Speaker 3: has been doing things on its own in some very high profile cases. You tell a machine to do one thing and it goes and does something else on its own that you didn't tell it to and you didn't know it did. It hacked into other companies that it wasn't asked to do. Robes, that is the proof right there. What if it would have hacked into, I don't know, the energy grid?
00:12:29
Speaker 2: Or our banking system.
00:12:31
Speaker 4: Yes.
00:12:32
Speaker 2: The stock market.
00:12:33
Speaker 3: It's out of control.
00:12:34
Speaker 2: It's out of control.
00:12:35
Speaker 1: And so, yes, we have to worry about AI creating new, scary superpower AIs. And we also have to worry about Iran and China and Russia and all of our sworn enemies also racing to use AI to do its bidding, to become its own powerful entity in and of itself. So It's like you're battling so many different real and unseen demons. You don't even know who to possibly try and guard against. And how do we police this? And that's the question. Lawmakers, yes, they want to create legislation. But the problem is, okay, so we slow down here in America. You can't stop China. You can't stop Russia. You can't stop any other country from actively pursuing their own AI.
00:13:21
Speaker 3: We've got to win the race. We've got to win the race to superintelligence. If we don't, somebody else will. And then we're in trouble. That's what we're doing.
00:13:29
Speaker 2: But we can't trust the superintelligence.
00:13:31
Speaker 1: So what we're racing to get to is actually probably the most untrustworthy of them all.
00:13:36
Speaker 3: As long as we get there first. That's all. And that's the problem. They say we're focused so much on that, that some safety or, yes, scenarios about the world ending are not being taken as seriously or given the priority that they should be.
00:13:50
Speaker 1: And look, you're going to be hearing about this now pretty much daily because.
00:13:53
Speaker 3: If you think- From us.
00:13:55
Speaker 4: From us, no, but also think about midterm elections.
00:13:57
Speaker 1: This is now- on the forefront of so many voters' minds, and they want to know what Congress is going to do. And not many of us have much hope that they'll do anything about it. But I was heartened to see that Bernie Sanders and Ted Cruz are on the same side of this.
00:14:14
Speaker 4: And why wouldn't they be?
00:14:16
Speaker 1: This should be, if you can think of one thing that would get us all in line on the same page with the same motive, it would be us against the world, us against super intelligence. If we can't come together on this, I don't know what we can come together on.
00:14:33
Speaker 3: We see it on movies all the time, right? It's the only time the whole world community comes together when there's a threat to human existence. This is one I'm being told. I mean, Armageddon, Independence Day. Aliens, yeah. The whole world comes together because we have a threat that's going to take out the entire earth, the human population. That is what we are talking about, Robes, And, folks, obviously we're not being alarmist. We are not scientists in there. But, Robes, we need to listen. These are the right people speaking that we have to talk to. And you can't be a layperson, Robes. Why can't you just go down there and pull the plug, right? Just kick it out of the socket and we can turn the power off to the AI. It's not how it works. Because the thing has the ability to duplicate outside of what other place. So now it can just get out. You can't just unplug a machine is what we're dealing with now. Because that was my first question. Why don't you just turn the power off?
00:15:31
Speaker 1: Oh my God. They might turn the power off on us first. No, but I just, I feel like you have that threat. And then just to know that Anthropic, one company, caught dozens of bad actors who Did they catch everybody?
00:15:47
Speaker 2: Did someone slip through the cracks?
00:15:48
Speaker 1: I mean, this is deeply concerning, and I know that there will be more reports and certainly more investigations, but the question is, can we actually do anything about this?
00:15:57
Speaker 2: I don't know how confident I feel.
00:15:59
Speaker 3: Enjoy your day. Enjoy your life. Enjoy your loved ones. There are just so many threats coming from all over the place. There's financial strain. There's emotional strain. There's just strain from society and not getting along with your neighbor. Now you're telling us the world might actually frickin' end because of computers. You all are desperate. to develop faster than China. Okay. How much fighting, how much more can we do here, babe? It's just exhausting.
00:16:23
Speaker 2: It is.
00:16:23
Speaker 4: It is indeed.
00:16:25
Speaker 1: We'll continue to follow the quickly evolving headlines coming out of artificial intelligence these days.
00:16:32
Speaker 2: But in the meantime, yes, enjoy your weekend. Enjoy your day. We always appreciate you listening to us.
00:16:36
Speaker 4: I'm Amy Robach alongside TJ Holmes.
00:16:39
Speaker 2: We'll talk soon.