Updated
Updated · TechCrunch · Aug 12
Hinton, Li and Ng Back Open AI at Ai4, Urge Regulation Over Gatekeepers
Updated
Updated · TechCrunch · Aug 12

Hinton, Li and Ng Back Open AI at Ai4, Urge Regulation Over Gatekeepers

3 articles · Updated · TechCrunch · Aug 12

Summary

  • At Ai4 in Las Vegas, Geoffrey Hinton, Fei-Fei Li and Andrew Ng all argued AI should remain meaningfully open, warning that a handful of dominant labs could end up controlling access, innovation and market power.
  • Ng framed openness as a competition issue, saying multiple providers are needed to prevent gatekeepers and to keep U.S. AI competitive against lower-cost open-weight models from China.
  • Hinton drew a sharp line between open-source code and open-weight models, saying released weights can be repurposed for harms such as cyberattacks, but added that widespread open-weight access is already irreversible.
  • Li rejected a simple open-versus-closed binary, arguing different layers of AI should have different levels of openness—like other scientific fields—and citing public-private models such as the Human Genome Project.
  • All 3 converged on one point: regulation is needed to steer AI toward public benefit rather than leaving key decisions to a few powerful companies or executives.

Insights

Are tech giants using AI safety fears to build an unbreakable monopoly, or is the danger truly that severe?
Could the blueprint for safely regulating tomorrow's superintelligence actually be hidden in the history of the Human Genome Project?