Science & technology | AI benchmarking
How to find the smartest AI
Developers are building fiendish tests only the best models can pass
THE DIZZYING array of letters splattered across the page of one of Jonathan Roberts’s visual-reasoning questions resembles a word search assembled by a sadist. Test-takers aren’t merely tasked with finding the hidden words in the image, but with spotting a question written in the shape of a star and then answering that in turn (see below).
Already have an account?Log in
Continue with a free trial
Get full access to global reporting that goes beyond the news cycle—for free
Or create a free account to unlock just this article
Create account Explore more
The Economist today
Handpicked stories, in your inbox
A daily newsletter with the best of our journalism
Climate change will hurt the richest farmers—and the poorest
Even with realistic adaptation, crop yields will fall as temperatures rise
Are China’s universities really the best in the world?
Nature’s prestigious index says yes
Meet the moths that use the stars to find their way
The skill was previously thought unique to humans and certain birds
The world needs to understand the deep oceans better
Otherwise it cannot protect them properly
Is the “manopause” real?
If it is, it is nothing like the menopause
A routine test for fetal abnormalities could improve a mother’s health
Studies show these can help detect pre-eclampsia and predict preterm births
