Demis Hassabis Claims AGI Will Outscale the Industrial Revolution by Tenfold

AGI: A Ten‑Fold Leap Over the Industrial Revolution

Demis Hassabis, Nobel laureate and head of DeepMind, announced that artificial general intelligence (AGI) is on the verge of realization and will bring changes "ten times the scale and speed of the Industrial Revolution." He predicts that drug discovery, clean energy, and advanced materials will accelerate dramatically, eventually making resources no longer a constraint on human progress.

Proposed Regulatory Framework

To manage this rapid transformation, Hassabis suggests establishing a new "standards body" modeled after the financial regulator FINRA. The body would oversee a "pre‑review" period of up to 30 days for any frontier AI lab before a model is released, and in extreme cases could coordinate globally to slow development.

Beyond Technical Challenges

He warns that even if technical hurdles are solved, deeper economic and philosophical questions will arise in a post‑scarcity world: what new economic models will ensure shared prosperity, what values should guide humanity, and how the human condition itself will change.

"A Framework for Frontier AI and the Dawning of a New Age"

In the translated article, Hassabis describes AGI as a singular moment comparable to discovering fire. He emphasizes that AGI’s impact will be unprecedented, potentially ten times the Industrial Revolution, and will usher in a new era of abundance.

Frontier Challenges

While AI already yields tangible benefits, realizing its full promise requires careful navigation of risks. Current frontier models pose cybersecurity threats, and as capabilities grow, risks such as nuclear and biological misuse may emerge. The rapid commercial and geopolitical competition accelerates progress faster than our understanding of the technology.

Policy Recommendations

Hassabis advocates for "cautious optimism": public policy should promote innovation while incentivizing responsibility and safety, foster international collaboration on critical security issues, and encourage thoughtful deployment of AI for societal benefit.

The Frontier AI Standards Body

The proposed standards body would:

Develop dynamic, flexible, and rigorous testing regimes for frontier AI models.

Be led by the United States initially, with a board comprising top technical experts and open‑source community representatives.

Secure substantial funding, primarily from industry, to attract world‑class talent and provide the compute needed for large‑scale testing.

Define a set of benchmark tests; models meeting the thresholds would be classified as "Frontier‑class" and subject to periodic updates.

Require recognized "Frontier Labs" to adopt best practices such as publishing model cards, maintaining strong internal cybersecurity, conducting background checks on key personnel, and allocating resources for safety research.

Implement a voluntary 30‑day pre‑review of models before public release; successful evaluation could become mandatory for deployment in the U.S. market.

Include rigorous scientific assessments of security, bio‑risk, and other high‑risk capabilities, with special tests for agentic AI to detect attempts to bypass safeguards.

Update tests regularly (e.g., quarterly), retire saturated benchmarks, and maintain held‑out test sets to prevent over‑fitting.

Collaborate with governments to create a third‑party audit ecosystem and co‑develop new benchmarks.

The framework aims to keep pace with rapid AI advances, tighten controls as risks are identified, and, if necessary, coordinate a global slowdown of development.

Recognition and Scope

Achieving "Frontier Lab" status would confer prestige and grant access to the framework, which applies to any frontier‑class model regardless of origin (nation, open‑source, or closed‑source). Non‑frontier models from startups or academia would be exempt.

Vision for a Shared Future

Hassabis concludes that AGI can become the ultimate tool for scientific and medical breakthroughs, driving massive productivity gains and economic growth. Realizing this potential requires a globally shared framework, the strictest scientific methods, and the collaboration of the brightest minds to address both technical and societal challenges.