AI has a stupid secret: we’re still not sure how to test for human levels of intelligence

By Andrew Rogoyski, Innovation Director - Surrey Institute of People-Centred AI, University of Surrey

Two of San Francisco’s leading players in artificial intelligence have challenged the public to come up with questions capable of testing the capabilities of large language models (LLMs) like Google Gemini and OpenAI’s o1. Scale AI, which specialises in preparing the vast tracts of data on which the LLMs are trained, teamed up with the Center for AI Safety (CAIS) to launch the initiative, Humanity’s Last Exam.

Featuring prizes of US$5,000 (£3,800) for those who come up with the top 50 questions selected for the…

Read complete article

~	Dockworkers pause strike after Biden administration’s appeal to patriotism hits the mark

~	A year of escalating conflict in the Middle East has ushered in a new era of regional displacement

~	Malaysia Deports Cambodian Worker for Critical Comments

~	‘Orphanage city’ helps children in Gaza as the war grinds on

~	UN rights council examines nuclear legacy consequences in Marshall Islands

~	Rights expert urges universities to respect pro-Palestinian protests

~	UK hands Chagos Islands to Mauritius, marking the end of a longstanding sovereignty dispute

~	Joker: Folie à Deux is relentlessly claustrophobic – and castigates its fans at every turn

~	Bird flu patient had no apparent contact with animals – and there’s still no evidence of sustained human-to-human spread

~	Can China’s stimulus blitz fix its flagging economy?