DeepMind founder Demis Hassabis argues that artificial general intelligence (AGI) may be realized in the near future and urges the United States to create an independent frontier‑AI standards body. He proposes a 30‑day pre‑release review to evaluate advanced models for security, bio‑risk and national‑security implications.

मुख्य बिंदु (Key Takeaways)

  • AGI could arrive within the next few years, according to Hassabis.
  • Proposal for an independent US frontier‑AI standards body via public‑private partnership.
  • 30‑day pre‑release assessment focused on cybersecurity, biological threats and national security.

Demis Hassabis, founder and CEO of DeepMind, outlined in a detailed blog post that artificial general intelligence (AGI) – systems with human‑level cognition – is likely to materialise within a few years. He stressed that without clear, enforceable safety standards, the societal impact of such technology could be both transformative and perilous.

Why a New Regulatory Framework is Needed

The US government has recently intensified scrutiny of frontier AI models such as Anthropic’s Mythos and OpenAI’s GPT‑5.6. In response, Hassabis recommends establishing an independent regulator, modeled after the Financial Industry Regulatory Authority (FINRA), to assess the most advanced AI systems before they reach the public. He argues that existing voluntary reviews are too lax, while traditional regulatory processes are too slow for the rapid pace of AI innovation.

How the 30‑Day Pre‑Release Review Would Work

Under the proposed framework, developers would voluntarily submit their “frontier lab” models for a mandatory 30‑day evaluation prior to launch. During this window, the regulator would conduct rigorous tests for cybersecurity vulnerabilities, biological misuse potential, national‑security risks, and any attempts by the model to bypass safety controls. If successful, the process could become a de‑facto requirement for any advanced AI released in the United States.

Potential Benefits and Risks of AGI

Hassabis likens AGI to the discovery of electricity or fire – a technology that could accelerate drug discovery, clean‑energy breakthroughs, and advanced material science by an order of magnitude, delivering economic growth ten times faster than the Industrial Revolution. Yet he warns that the speed of AI capability growth outpaces society’s understanding of its hazards, especially in cyber‑warfare, bio‑weaponisation, and autonomous weapons.

International Implications and the Path Forward

Although the proposal targets the US, Hassabis envisions it becoming a template for global standards. He writes, “We must use this precious window before AGI arrives to shape this technology for the benefit of all humanity.” By aligning innovation incentives with robust safety mandates, the framework could foster international collaboration on AI risk mitigation while preserving the transformative promise of AGI.