AI language models killed the Turing test: do we even need a replacement?

"Let's figure out the kind of AI we want, and test for those things instead,"

"this march towards AGI, I think we're really limiting our imaginations about what kind of systems we might want - and, more importantly, what kinds of systems we really don't want - in our society"

"The idea of AGI might not even be the right goal, at least not now,"

The Turing test uses short, text-based conversations in which a judge interacts with a human or a machine; a machine passes by convincing the judge it is human. Contemporary models often meet such imitation benchmarks, prompting debate over the test's ongoing value as a progress metric toward artificial general intelligence (AGI). Critics argue that AGI is an ambiguous goal and that focusing on imitation constrains imagination about desirable and undesirable systems. Proposals favor abandoning the Turing test in favor of assessments that evaluate AI safety and specific capabilities that provide clear public benefit. A commemorative gathering marking the imitation game's 75th anniversary attracted wide public interest.

Read at Nature