A robot ranking looks reassuringly simple until the arms, training data and test rules come into view. The same software can face quite different examinations in different laboratories. Playground’s new index makes a useful attempt to join those results, while exposing how much judgement sits inside a single score.
This article continues for subscribers. Subscribe to read the full analysis.
SectorFrame membership
Read the full story
Independent reporting for people putting AI to work. Sign in to check your access, or explore a subscription.