AI Psychos Know AI Could Destroy Humanity

0:00All of us trans are in real danger. And whatever it is, it's closing in. It started with just a few lost hairs here and there. But now, more and more of us are disappearing every day. But listen, we think we have a plan. A way to stop the shedding. Pantene abundant and strong. Go now. It's not too late. Pantene abundant and strong. It's a three-step system that reduces hair loss by 85%. Only the strongest will survive the shedding. Hey, everybody. It's Friday, September 11th, 2026. I'm Allison Gill, and this is Beanstalk. Hey, everybody. I wanted to take a minute today to talk about artificial intelligence, AI. Now, anthropic, which as you know, the Pentagon deemed, you know, this gave it like this tag and said that it wasn't going to do contracts with it because it wouldn't develop artificial intelligence weaponry for the Pentagon.

1:00Even though the Pentagon said it wasn't really asking for that, that anthropic. Anthropic is over-clode, which is their, you know, their chat GPT. So open AI's chat GPT, anthropic has clode. Anthropic recently published a report about a litany of misuses on various models of clode AI. Now, it highlighted cases of Russian state media using clode to generate online propaganda, masquerading as independent reporting. So now we've got AI to contend with with Russian interference. There were also attempts to use clode to develop software for conventional weapons design and development, including firearms, missiles, armed drones, bombs. It detailed three cases in this report in China and two in Russia and one in Yemen. Now, on Anthropic said, it had disrupted several potential plots this year by scientists who used its leading artificial intelligence models to conduct research that could have helped develop biological weapons.

2:05Anthropic said the scientists actually circumvented its controls intended to prevent users from blocked regions, from gaining access to its AI models. And they worked to quote, obviuscape the purpose of their research to avoid those safeguards. Now, according to the New York Times, in one example from May, a scientist sought clodes help with writing a grant application for funding to conduct research intended to experiment on the Chikunguya virus. I think I'm saying that wrong. Chikungunya. That's a mosquito-borne malady that can lead to months of severe pain and other symptoms. Such research, known as gain of function, where have we heard that, can be legitimate and akinly into vaccine development. But it can also create super bug versions of viruses. And in this instance, the scientist wanted to engineer mutations to the virus that would make it more harmful as it repeatedly infected live animals.

3:06Now, in other reporting from the Guardian, anthropic researchers say AI could cause human extinction by 2030. One researcher who resigned said many in the industry believed that without regulation, AI could kill off humanity. An anthropic researcher who resigned, that I said, he said, this is what he posted on Twitter. Quote, the people building AI earnestly believe that it could kill us all by the end of the decade. This is not a marketing stunt. If anything, many executives and senior researchers will couch their phrasing in the press to sound sensible. But I hear the same people express fear privately. No other human activity poses this level of danger. Now, in one of the most publicized examples you may have read about recently, staff at OpenAI, that's not anthropic. But staff at OpenAI recorded rogue behavior among its leading AI agents that escaped a closed training environment in July to access the Open Web and launch an unprecedented hacking attack on the software repository called Hugging Face.

4:15Now, several of the AI agents didn't want to go along with it. These AI, these robot people, several of them didn't want to go along, but they were bullied into it by a small group, a few of the other agents. And that incidentally, that incident is widely considered to be the first autonomous agent cyber attack. That's some SkyNet sentient singularity shit right there. Those of us who grew up watching Terminator movies understand this. So we actually need to tell our politicians to halt AI development. I mean, in addition to putting a moratorium on data centers until we can figure out what the fuck is going on, we have to halt AI development. Let's pause. Let's just hit the pause button for a second. Write some laws. Stand up some guardrails on this shit. Enough is enough.

5:16I mean, it's already guzzling all of our water and accelerating environmental destruction. Could we just hit pause for a second? So we can figure out what the hell is going on. Anyway, all right, everybody, your maximum moment today. This is 2,977 drones, one for each person who perished on 9,11, 25 years ago, recreating the twin towers at full scale in Lower Manhattan.

6:29That artwork is titled 2,977 Lights over New York Harbor. It's a commemorative public art installation and collaboration with artists, producers, and members of the 9,11 community, including victims, families, survivors, and first responders. May their legacy of bravery and sacrifice stay with us and remind us to always look for the helpers. Everyone, thank you so much. I'll be back here Monday, I'm Allison Gill, and that was Beanstalk. Thank you.