Middle Powers Seek To Enforce AI Safety

At the New Delhi AI Impact Summit last week, middle powers including India and Canada warned that the US, China and major tech firms will not meaningfully restrain AI, while the New Delhi Declaration remained non-binding. The columnist urges these countries to unite and condition market access on measurable safety measures—such as training-data and energy disclosures, independent evaluations, and incident reporting—amid about $650 billion in expected US AI investment this year.