US Withholds Support for AI Safety Report Ahead of India’s AI Impact Summit

The United States’ decision to withhold support from a major new international AI safety report has cast a political shadow over next week’s India–AI Impact Summit in New Delhi, where more than 20 heads of state and delegations from over 45 countries were expected to debate how to govern rapidly advancing AI systems orfonline +1. The report, released 3 February and backed by around 30 governments and organisations, warned that current safeguards for powerful general‑purpose AI remain “insufficient” even as evidence of cyber, biological and other risks has “grown substantially” crowell +1.

Why Washington Broke Ranks on the AI Safety Report

The second International AI Safety Report, chaired by Turing Award winner Yoshua Bengio and drafted with input from about 100 experts, was designed as a shared evidence base for frontier‑AI risks, from automated hacking to assistance with biological or chemical threats crowell +1. While the European Union, United Kingdom, China, Saudi Arabia and others participated in its advisory panel or drafting process, the U.S. — which had engaged with earlier drafts — declined to formally back the final text, TIME reported on 13 February ndtv +1.

Washington has not publicly offered a detailed explanation, but analysts framed the move as part of a broader reluctance to be bound, even symbolically, by multilateral language that could constrain U.S. tech firms or pre‑empt domestic rulemaking ndtv. Bengio argued that “the greater the consensus around the world, the better,” underscoring concerns that a visible U.S. absence might weaken efforts to align national standards on testing and restricting high‑risk AI systems ndtv.

India’s AI Impact Summit Becomes the New Battleground

The diplomatic rift landed days before India hosts the India–AI Impact Summit from 16–20 February at New Delhi’s Bharat Mandapam, billed as the first major global forum focused on AI’s “impact” and deployment at scale orfonline +1. India’s government said the summit would be anchored in three “sutras” — People, Planet and Progress — and has invited leaders including France’s Emmanuel Macron, Brazil’s Luiz Inácio Lula da Silva and the UAE’s Mohamed bin Zayed Al Nahyan, alongside tech chiefs from companies such as Google, OpenAI and Nvidia datainnovation +1.

Indian officials have pitched the event as a chance to champion “democratisation of AI” and applications for healthcare, agriculture and public services, while outside experts urged New Delhi to keep safety “front and center” despite the summit’s development‑heavy branding msn +1. With the U.S. distancing itself from the safety report, negotiations in Delhi are expected to highlight competing governance models: Europe’s binding regulation via the EU AI Act, looser U.S. experimentation, and India’s bid to position itself as a bridge between advanced and developing economies reuters.

The Bigger Picture

The split over the AI safety report turned what might have been a largely technical document into a marker of geopolitical fault lines at the very moment governments were seeking common ground in Delhi. For countries worried about both the economic stakes of AI and the security risks flagged in the report, the summit will test whether broad agreement on scientific evidence can translate into aligned — or at least interoperable — rules without Washington’s full endorsement.

orfonline India AI Impact Summit 2026 / IndiaAI Mission
ndtv TIME, “U.S. Withholds Support From Global AI Safety Report”
crowell International AI Safety Report 2026 (full PDF)
tipranks Inside Privacy summary of the report
timesofindia International AI Safety Report 2026 – methodology and contributors
time International AI Safety Report – Wikipedia entry
ciol India AI Impact Summit 2026 – about and schedule pages
datainnovation Indian Ministry of External Affairs press release on world leaders attending
atlanticcouncil Indian and international tech press listings of CEO attendees
msn DD News / India AI Impact Summit 2026 preview
americanprogress Atlantic Council commentary on safety at the summit
reuters EU AI Act trackers and analysis