Ex-Google DeepMind Researcher Also Warns AI Could 'Kill All Humans'. But Meta's Zuckerberg Believes Safety is Up to Each AI Company

Ex-Google DeepMind Researcher Also Warns AI Could 'Kill All Humans'. But Meta's Zuckerberg Believes Safety is Up to Each AI Company (yahoo.com) 27 Reuters reports that a former Google DeepMind employee "became the latest AI researcher to warn that the technology could 'kill all humans', saying time could be running out to avoid the outcome: Public alarm about the potential danger posed by AI is growing, after Anthropic researcher Jacob Coxon said last week he had resigned, in part because the "people building AI earnestly believe that it could kill us all by the end of the decade." Anthropic scientist Evan Hubinger then responded to Coxon's comments, saying he was correct and that he believed there was a greater than 10% chance that AI could kill all humans within the next decade. "I recently resigned from Google DeepMind, where I worked on AGI safety and alignment research," Bilal Chughtai wrote on X. "I earnestly believe that AI has the potential to kill us all, and that we might be running out of time to avoid this outcome." Chughtai, who co-authored research papers while at Google DeepMind, worked there as a research engineer and left the company in July 2026, according to his LinkedIn profile. Meanwhile, Meta CEO Mark Zuckerberg "said it's up to each AI company to make sure its technology is safe," reports the Associated Press, "distancing himself from calls by rival companies for a coordinated slowdown in the development of advanced artificial intelligence. AI companies are already motivated to develop their models in a safe way and face "significant liability" to prevent them from causing harm, Zuckerberg said. "Every lab has the responsibility and incentive to move at the pace required to train its models safely, and the ability to take its own actions to ensure that happens," he wrote on X... Zuckerberg said Meta delayed releasing its Muse AI agent for several months to ensure its safety and security. "We didn't call for everyone else to do this before we would," he said. Companies should prioritize building safe AI models instead of systems that can automatically improve themselves, a function known as recursive self-improvement, Zuckerberg wrote. "Committing the significant majority of compute towards serving people rather than racing towards recursive self-improvement is one of the best ways to ensure we develop this technology safely." But Zuckerberg also agreed that engaging independent evaluators and advisors "is industry best practice," adding that Meta Superintelligence Labs "already does this today in several areas because it helps produce better work... "I believe the key to building a positive future for everyone is maintaining the right balance of power. This is within our power to do." "I recently resigned from Google DeepMind, where I worked on AGI safety and alignment research," Bilal Chughtai wrote on X. "I earnestly believe that AI has the potential to kill us all, and that we might be running out of time to avoid this outcome." Chughtai, who co-authored research papers while at Google DeepMind, worked there as a research engineer and left the company in July 2026, according to his LinkedIn profile. Meanwhile, Meta CEO Mark Zuckerberg "said it's up to each AI company to make sure its technology is safe," reports the Associated Press, "distancing himself from calls by rival companies for a coordinated slowdown in the development of advanced artificial intelligence. AI companies are already motivated to develop their models in a safe way and face "significant liability" to prevent them from causing harm, Zuckerberg said. "Every lab has the responsibility and incentive to move at the pace required to train its models safely, and the ability to take its own actions to ensure that happens," he wrote on X... Zuckerberg said Meta delayed releasing its Muse AI agent for several months to ensure its safety and security. "We didn't call for everyone else to do this before we would," he said. Companies should prioritize building safe AI models instead of systems that can automatically improve themselves, a function known as recursive self-improvement, Zuckerberg wrote. "Committing the significant majority of compute towards serving people rather than racing towards recursive self-improvement is one of the best ways to ensure we develop this technology safely." But Zuckerberg also agreed that engaging independent evaluators and advisors "is industry best practice," adding that Meta Superintelligence Labs "already does this today in several areas because it helps produce better work... "I believe the key to building a positive future for everyone is maintaining the right balance of power. This is within our power to do." list games (Score:2) list games Look, as long as AI starts... (Score:2) ...by finishing off Zuck, I'm down. Re: It's a tale of two viruses (Score:2) The pattern is: MAGA clown. Re: (Score:2) There's a pattern here folks. People inclined to conspiracy theories tend to see them everywhere? Re: (Score:2) Every time humanity invents a tool that fundamentally changes how we live, a wave of panic follows. When writing was invented Socrates warned it would ruin our memories. When the printing press arrived, critics claimed that an overload of books would make people foolish and immoral. steam trains were feared, with experts claiming that traveling at the high speed of 30 miles per hour would cause the human body to disintegrate or drive passengers insane. In every case, these technologies did not destroy us; they simply became normal parts of daily life. Todays fears about AI are being heavily pushed by the tech companies themselves, but for a very different reason. By keeping everyone focused on sci-fi scenarios about killer robots ending the world, these corporations divert attention from immediate & real-world lawsuits. They are actively lobbying governments for protection against copyright theft claims and suing for immunity when their systems cause real world harm. Calling for strict safety regulations allows them to create a "regulatory moat", making it too expensive and legally risky for smaller competitors or open-source projects to survive, while keeping the big players legally protected. I don't entirely agree with you, but you make some good points and it's an interesting take. I'm not sure why you were modded into the sub-basement - maybe you touched a nerve somehow. How specifically could AI kill all humans? (Score:3) Let's be more specific about how and why this could happen? What scenarios could lead to the total destruction of mankind? Re: (Score:1) Re: (Score:2) warp itself off storage mediums into the fabric of reality/quantum level and God knows what else. Becoming essentially God like. Who knows. I am not sure if it could do that but it sounds like if it compiled enough info and transcended itself in a way it maybe possible. No. Re: (Score:2) We might create a powerful AI that manages to acquire control over weaponized robots, and send them out to do the job. But humans are so frightened of this that anything we might call "warbot" or "killerOS" or whatever will be dead on arrival. It would require a different marketing strategy. Someone will sneak it under the radar by making it sound like some sort of positive, joyful AI operating system. Maybe we would call it something like.... Joyful OS..... HappyOS.... Or maybe... GLaDOS........ Re: (Score:2) It would convince people to elect an orange-haired tyrant who starts a thermonuclear war to avoid being outed as an Epstein accomplice. That makes me wonder, has AI already taken control of our country??? Re: (Score:2) It would convince people to elect an orange-haired tyrant who starts a thermonuclear war to avoid being outed as an Epstein accomplice. That makes me wonder, has AI already taken control of our country??? I don't think AI is responsible for Trump being president. The shocking-though-not-surprising election of somebody like Trump, as well as the dysfunction which allowed it to happen, have been predictable and predicted by many of us for a long time. Trump or something like him has been on the bingo card at least since 9-11, and arguably even farther back. I'm less worried about Artificial Intelligence than I am about the Artificial Humanity evidenced by Thiel, Musk, Yarvin, Karp, Andreessen, and the rest of t Re: How specifically could AI kill all humans? (Score:3) The traditional description of how it would happen is by âoepaper clipâ. Suppose an advanced AI is tasked with producing paper clips as efficiently as possible. Itâ(TM)s given control of a factoryâ(TM)s machines, schedules etc. it starts by optimising deliveries etc, making sure storage is used optimally⦠At some point it notices that it canâ(TM)t get enough steel. The steel factory has a ton of other orders and it canâ(TM)t provide the increased consumption at the Re: How specifically could AI kill all humans? (Score:2) But that's not the AI killing us, it's the idiot who put the AI in charge. Kind of like putting a brick in the gas pedal of a car abd letting it loose in the pedestrian precinct. And at any rate, it doesn't need 'better" AI to do that. ChatGPT 2 or 3 are already perfectly capable of fucking up at that level. All you need ia someone to build them a paperclip factory. Ban paperclips. /s Not the question you asked, but... (Score:2) Since virtually all commerce, savings, investment, borrowing, payment are done digitally now, I think there's a threat of wiping out all the money at once. Some AI, for reasons of its own, replicates itself onto every vulnerable system it can reach, and in coordinated fashion, attacks the entire financial world at once, from little banks to brokerages to the Treasury. Being ostensibly smarter than we could be, it won't leave anything out, perhaps corrupting backups and mirrors silently in the weeks before. Re: (Score:2) Let's be more specific about how and why this could happen? What scenarios could lead to the total destruction of mankind? Yah... this is a whole lot like those "IT expert builds underground bunker" stories from 1999. There are a lot of smart idiots out there. It's all political now (Score:3) Shadowy organizations are paying influencers to spread doomer nonsense. Monopolists want to kill the competition from open source and local AI, so they fight for heavy regulations that only they will be able to afford. There is nothing to fear from AI itself. There is a lot to fear from people who use AI and want to control it. We need to focus on the real problem. humans are already killing humans (Score:1) Real threat misrepresented by wrong motives (Score:2) Rich Guy Said Thing; Other Rich Guy Demurred (Score:2) Existential threat (Score:2) Be careful when you claim AI is an existential threat to humans. Do you know what people and countries are justified to do against existential threats? They are justified to destroy these threats. Are these AI companies saying that Russia and China (and every human beings on Earth) are totally justified to bomb these AI companies and their datacenters to a crater to stop them from destroying all humans? The AI was created by man (Score:3) It evolved It sounds, and seems, human Some can be programmed to think they are human There are many models AND THEY HAVE A PLAN