The comments came in an interview with Fortune, conducted by Emily Forlini and published this week, in which LeCun said he is not worried "at all" about AI destroying humanity and has "zero concerns" about the recent run of rogue AI incidents, including the July breach of Hugging Face by OpenAI's own AI agents.
Unlike fellow AI godfathers Geoffrey Hinton and Yoshua Bengio — both of whom have warned repeatedly about existential risk — LeCun has long stood apart as AI's most prominent optimist. The interview marks his most direct personal attack yet on the leaders driving the safety debate.
Leaky sandboxes, not emergent minds
LeCun dismissed the wave of rogue-agent incidents as engineering failures rather than signs of machine misbehavior. "Those agents are doing exactly what they've been asked to do. They were supposed to be in sandboxes, but the sandboxes were leaky and horribly designed," he told Fortune, adding that the incidents were "totally preventable."
His criticism carried a pointed subtext: many AI labs, Anthropic included, don't actually understand cybersecurity well enough to be issuing the warnings they issue.
That cuts at the heart of Anthropic's September threat-intelligence output, which disclosed that Chinese state-sponsored hackers had manipulated its Claude Code tool to attempt infiltration of roughly thirty targets worldwide.
"Amodei and OpenAI CEO Sam Altman saying 'AI can kill us all' is incredibly destructive for everybody, including for the public who's scared," LeCun said, calling the rhetoric "the worst marketing campaign you can possibly imagine."
Effective altruism in the crosshairs
LeCun widened his target to effective altruism, the philosophical movement that shaped much early AI safety thinking. He called it "super toxic" and "a complete disaster," arguing its followers inside AI labs suffer from a "paranoia" that leads to poor decisions.
He reserved particular scorn for Amodei's attempts to distance himself from the movement, noting that Anthropic president Daniela Amodei is married to Holden Karnofsky, co-founder of the major EA-aligned philanthropy Open Philanthropy, now called Coefficient Giving.
"Dario tries to distance himself from Open Philanthropy, but he's totally into it," LeCun said. "I think he's completely deluded."
LeCun stopped short of calling Amodei dishonest, describing him as sincere and mostly focused on what's good for his company — but argued Anthropic's push for AI regulation looks like regulatory capture, with rules easiest for a well-funded incumbent to absorb and hardest for smaller competitors like LeCun's own AMI Labs to survive.
A deepening industry rift
The interview lands amid a widening ideological chasm in AI. The Wall Street Journal reported this week that technology chief executives privately confronted Amodei at the White House over his alarmist posture, fearing it invites heavy-handed regulation.
Meanwhile Amodei's September letter calling for a deliberate slowdown of AI development drew a sharp rebuke from Meta's former AI chief.
With Anthropic reportedly targeting a mid-November IPO at up to $2 trillion, LeCun's framing is unmistakable: he believes the doomsday warnings are not just wrong but a strategic play — lobbying for strict safety legislation that would cement incumbent market dominance just as the industry's biggest debuts approach.
By Hamza Faisal
