Verity Harding warns that calling AI an arms race could backfire

In short: A former DeepMind policy lead says the popular “AI arms race” story can steer governments toward rivalry and away from cooperation.

What's going on

Verity Harding, who previously briefed world leaders on artificial intelligence at Google DeepMind, argues that the way people talk about AI is shaping real government decisions. She says AI research once leaned more on international cooperation, but the tone has shifted toward competition between companies and between countries, especially the US and China.

Harding curated a new essay anthology called Reframing the AI Arms Race. Contributors include people from politics and academia, such as historian Lawrence Freedman and Japanese politician Taro Kono. Their main point is simple: words matter, because they set the mood for policy. Harding says the “arms race” label is appealing because it sounds clear, but it can also narrow thinking.

She links the rise of this framing to recent events that made geopolitics feel more tense, including the launch of ChatGPT in late 2022, the pandemic, and the war in Ukraine. Harding also points to US government rhetoric and moves that treat AI more like a national asset to control, including export controls, which are rules that limit who can buy certain technology (like putting a lock on who can get a powerful tool).

What to watch

Harding warns that if AI is treated mainly as a weapon, countries may collaborate less on shared problems like security and health. She suggests a “middle powers coalition,” meaning countries like Canada, France, Japan, South Korea, India, and the UK working together so they are not forced to pick sides. The next signal to watch is whether more governments copy the arms race language and tighten restrictions, or whether they create more cross-border agreements that focus on safety and shared rules.

Source: Wired