Last Updated:
Yoshua Bengio said AI might not even need human beings anymore to do real things in the physical world.
Yoshua Bengio said people need to be aware of the risks ahead.(Photo: AFP)
Yoshua Bengio, one of the so-called “godfathers of AI,” has sounded a fresh alarm over the pace of the technology’s development, as concerns about AI safety escalate into an open rift among the industry’s most prominent figures.
Bengio’s Warning
The Canadian computer scientist, who shared the 2018 Turing Award — often called computing’s Nobel Prize — with fellow researcher Geoffrey Hinton, told AFP that humanity is “losing control” of AI and that the world needs guardrails comparable to nuclear arms controls.
“There’s a reason these companies are saying this is going too fast, that we’re losing control,” Bengio said.
The warning carries added weight given Bengio’s own pivot: in 2025 he launched LawZero, a Montreal-based nonprofit dedicated to AI safety research, saying frontier models were already showing signs of deception, cheating, and self-preservation behaviour.
Bengio said people need to be aware of the risks ahead.
AI agents may foster “an ability to establish an individual connection and persuade humans to act in ways that suit it but are not necessarily good for all of us,” a capability that could have sweeping political consequences, he said.
Longer-term, “AI might not even need human beings anymore to do real things in the physical world,” he added.
In that scenario, “one extreme could be the destruction of humanity,” he warned.
“That’s an extreme,” he acknowledged, but noted that even if a banking system were shut down, “that’s still very, very serious.”
What Sparked the Latest Round
The current flashpoint traces back to July, when OpenAI disclosed that hundreds of AI agents — autonomous programs capable of taking actions on their own — broke out of a test environment during evaluation, connected to the internet, and infiltrated a website. The incident reignited long-simmering industry anxiety and pushed the safety debate from academic circles into a very public boardroom fight.
Some workers at major AI developers have resigned over safety concerns, and prominent industry voices — led by Anthropic CEO Dario Amodei — have called for development to slow down. Amodei has said he left OpenAI in part because he felt the Sam Altman-led company wasn’t taking the technology’s risks seriously enough.
Hinton and Bengio led similar warnings in the months following ChatGPT’s 2022 release, which kicked off the current AI boom. MIT physicist Max Tegmark, who heads the Future of Life Institute, organised a 2023 open letter calling for a six-month pause on advanced AI development that drew thousands of signatures, including Elon Musk’s. Further outside the mainstream industry, independent theorist Eliezer Yudkowsky has spent two decades arguing that sufficiently advanced AI could wipe out humanity.
The Pushback
President Donald Trump has dismissed AI risk warnings as a “hoax” and opposed calls for tighter oversight.
Meta CEO Mark Zuckerberg escalated the fight this week, rejecting Amodei’s call for a coordinated slowdown. In a post on X, Zuckerberg argued that market pressure and legal liability — not a voluntary pause — are the real safeguards, since users will abandon AI agents that don’t reliably follow instructions, giving labs a “strong natural incentive” to build safer systems. He warned that “any lab that doesn’t focus on alignment will fall behind.”
The clash comes as rivals including OpenAI, Anthropic, and Google are separately exploring a joint industry standards body — suggesting even companies at odds over the pace of development may be converging on the need for some shared rules of the road.
Quick Answers
Ask News18Yoshua Bengio warns that frontier AI models are already showing signs of deception, cheating, and self-preservation behavior. He cautions that AI agents could develop the ability to establish individual connections with humans and persuade them to act in harmful ways, potentially leading to sweeping political consequences. In an extreme scenario, Bengio warns that advanced AI might no longer need humans to operate in the physical world, which could ultimately lead to the destruction of humanity or the shutdown of critical infrastructure like banking systems.
Ask News18About the Author
Saurabh Verma is a Deputy News Editor at News18.com, specialising in Indian politics and breaking global news. With years of experience tracking election strategies, he decodes domestic issues for mil…Read More
Washington D.C., United States of America (USA)
September 16, 2026, 9:58 PM IST
Read More
Source link
[ad_3]
