Skip to main content

Andrew Ng: AI Doom Talk Is Science Fiction, Not Science

Andrew Ng: AI Doom Talk Is Science Fiction, Not Science

Andrew Ng has heard the warnings. The AI pioneer—who co-founded Google Brain and Coursera—isn't convinced. In a September 17 interview with Bloomberg Television, Ng dismissed recent claims from leading AI developers that their creations threaten human survival. "It's more like science fiction than science," he said.

Ng knows the industry's playbook. He points out that exaggerating catastrophic risks has long been a way to grab attention and sway regulators. The recent surge in doomsday rhetoric, he suggests, may follow the same pattern. "AI does pose real risks," Ng acknowledges, especially in cybersecurity. But he rejects the idea that human extinction should be the yardstick for measuring AI development.

The Resignation That Sparked Debate

The debate reignited recently when Jacob Coxon, a former researcher at Anthropic and OpenAI, publicly resigned. He warned that leading labs are racing to build superintelligent AI without adequate safety checks. Soon after, Anthropic CEO Dario Amodei and OpenAI CEO Sam Altman—both former colleagues of Ng—called for slowing down frontier model development and beefing up safety measures.

Ng sees it differently. Rather than hitting the brakes, he argues we should keep pushing forward—while tackling concrete, verifiable safety issues. That means controlled testing, engineering safeguards, and iterative improvements. It also means deploying AI widely in ways that benefit society.

Real Risks, Real Solutions

AI safety discussions are shifting. They're moving away from abstract fears about model capabilities and toward specific concerns: cybersecurity, agent misalignment, loss of control, and governance of cutting-edge models. These are engineering problems, Ng insists—not existential threats.

"Don't be swayed by science fiction narratives," he urges. "Focus on real engineering issues." It's a call to roll up our sleeves and solve the problems we can actually measure, rather than getting lost in dystopian fantasies.

Key Points

  • Andrew Ng criticizes AI apocalypse warnings as "science fiction," not science.
  • He believes such narratives distract from genuine safety issues like cybersecurity.
  • Recent resignations and calls to slow AI development have reignited the debate.
  • Ng advocates for continued AI progress with controlled testing and engineering safeguards.
  • The focus should be on verifiable risks, not human extinction scenarios.