Former DeepMind Director: US Nationalist AI Policy Leads to Catastrophe

Verity Harding, former director of DeepMind, stated in an interview with Wired: the nationalist course of the American government in the field of artificial intelligence is no longer a theoretical danger, but direct evidence that the worst-case scenario is beginning to materialize. In her conviction, the AI arms race is capable of turning into a catastrophe.

What Harding sees from inside

Harding worked at DeepMind — the flagship AI laboratory within Google, one of the most influential research organizations in the world. Over the years working at the intersection of technology and policy, she has formed a clear position: the main threat does not come from the technology itself, but from how states build policy around it.

When governments view AI primarily through the lens of national superiority, the familiar logic of arms race kicks in. "Defeating the opponent" becomes more important than "developing responsibly." Development speed displaces consideration of consequences. Secrecy triumphs over transparency.

"The nationalist approach of the American government to AI is evidence that the worst-case scenario is beginning to materialize,"

Harding told Wired.

Why AI nationalism is more dangerous than previous races

Geopolitical rivalry in the field of AI carries risks fundamentally different from previous technological races. Nuclear technology was created with high barriers and remained in the hands of states. AI capabilities spread differently: through open publications, commercial APIs accessible to millions of developers around the world.

AI nationalism creates several systemic problems at once:

• International exchange of information about risks and vulnerabilities is blocked due to competitive considerations
• The pace of development is artificially accelerated — outpacing researchers' ability to assess consequences
• Security researchers lose influence over priorities, which are increasingly dictated by geopolitics
• Global control mechanisms do not have time to form — each side expects that the rules will be established by itself

States that perceive AI as a geopolitical weapon are objectively not interested in safety measures capable of slowing down their progress. It is within this structural trap — according to Harding — that potential catastrophe is being born.

Technology versus geopolitics

Harding's position reflects a growing rift in the AI community. Some experts believe the main threat comes from the powerful models themselves — their capabilities and unpredictability. Others, like Harding, are convinced: technology itself is neutral; the problem is how political and military logics shape its development.

When leading technological powers build national AI strategies in the logic of "zero-sum," international forums on security turn not into platforms for cooperation, but into yet another field of geopolitical competition.

What this means

Harding's warning is part of a growing consensus among AI experts: while "AI nationalism" dominates over international cooperation in the field of safety, the risks of catastrophic failure remain open. Technology develops faster than the mechanisms for controlling it form — and it is precisely this gap that causes the greatest concerns.