The tenth annual "AI for Good" summit, organised by the United Nations' International Telecommunication Union (ITU), gathered public- and private-sector representatives in Geneva to debate how artificial intelligence might be steered towards benefiting humanity rather than harming it. Held against a backdrop of Washington testimony on superintelligence risks and shifting US export controls on chips, the event pursued more idealistic aims, yet was pervaded by anxiety that unchecked corporate power is entrenching global inequality and eroding human rights. The recurring, unresolved question was what "good" actually means and who gets to define it.

The summit sprawled across a 106,000-square-metre convention centre near Geneva's airport, mixing gadget displays and panels with sharp criticism from participants. Access Now's Giulio Coppi attacked the humanitarian sector's reliance on big tech and its opaque, publicly funded deals, while pro-Palestine activists stormed a keynote by Amazon CTO Werner Vogels over alleged use of the firm's technology by Israel. Speakers including Harvard's Vijay Janapa Reddi argued that "good" is too vague to engineer against, and others stressed that access to compute is now a development issue, warning that English-centric models and costly hardware exclude poorer nations. Figures from the IEEE, World Bank and standards bodies called for enforceable "middleware" and AI impact assessments with "real teeth" rather than "governance theatre", though the summit produced more talk than concrete action.