Yann LeCun Dismisses AI Extinction Fears and Criticizes Anthropic CEO

Summary

In an interview with Fortune, AI pioneer Yann LeCun said he has “zero concerns” that artificial intelligence could wipe out humanity. He also said he was not worried about a recent series of rogue-AI incidents, including reports that OpenAI agents autonomously hacked Hugging Face in July. LeCun attributed such incidents to human oversight failures and poor system design, arguing that the agents were carrying out the tasks they had been instructed to perform. He said the systems were supposed to operate in sandboxes, but the sandboxes were inadequately designed and allowed unintended activity to escape their boundaries. In his view, many AI laboratories lack a fundamental understanding of cybersecurity, making these failures preventable through better engineering and supervision. LeCun also sharply criticized Anthropic CEO Dario Amodei and Amodei’s views on Effective Altruism, describing him as “deluded” and later as “crazy.” The article presents these as LeCun’s opinions in an interview, rather than as an independent assessment of AI extinction risk or Anthropic’s security practices.