Former Google DeepMind policy leader Verity Harding warns that framing AI development as an arms race between superpowers risks closing off international cooperation essential for safe, beneficial technology. She argues the war metaphor—adopted after ChatGPT's launch in November 2022 amid pandemic and Ukraine war concerns—consolidates power in large labs and forces smaller nations to pick sides, and proposes a coalition of middle powers to preserve collaborative alternatives.
Summaries like this, in your inbox every morning.
Sign up free →What happened
Verity Harding, who previously briefed world leaders on AI while head of global public policy at Google DeepMind, has published an essay anthology titled Reframing the AI Arms Race arguing that war metaphors describing AI development are fundamentally misguided and shape policy dangerously.
Why it matters
Harding contends that casting AI as a lethal competition between superpowers closes the door to international cooperation needed for safe technology and fair benefit distribution, while smaller nations that accept this framing must choose sides—potentially against their own interests. The Trump administration's nationalist AI rhetoric and forced withdrawal of Anthropic's frontier model exemplify how the arms race narrative translates into real policy.
What to watch
Harding advocates for a middle powers coalition—potentially including Canada, France, Japan, South Korea, India, and the UK—to resist the binary superpower competition and preserve collaborative capacity on shared challenges like security and disease prevention.
The arms race metaphor for AI emerged quickly after ChatGPT's November 2022 launch, a moment when geopolitical tensions were already high due to the pandemic's end and Russia's invasion of Ukraine. Harding traces how the narrative shifted from international cooperation—the norm during her time at DeepMind between 2016 and 2020—to a civilizational battle between the West and China. Two forces drove this change: a sincere belief that AI technology was dangerous and should be controlled by democracies, and an anti-regulation lobby that benefited from China as a bogeyman, using the threat to discourage domestic oversight.
The stakes of this framing are material and structural. When nations accept the binary arms race logic, they surrender agency; smaller powers become "chess pieces on one side or the other." The major AI labs themselves have benefited from the narrative, Harding argues, because it concentrates power—only they claim the expertise and urgency to lead. The long-term trajectory Harding fears is excessive government control, centralized power over AI systems, and "vassal states" unable to pursue their own interests. Competition itself is normal and healthy, but the arms race language makes collaboration unthinkable just when it is most needed—on shared security, food, and health challenges where no single superpower can solve alone.
No discussion yet for this article
Get curated AI news from 200+ sources delivered daily to your inbox. Free to use.
Get Started FreeFree · takes 30 seconds · unsubscribe anytime
1 minute a day. The AI essentials.
200+ sources · Email / LINE / Slack
