How to tell whether AI is behaviourally human-like
Year: 2026
Type: preprint
Abstract: AI systems are increasingly described as behaving like humans. But passing as human in conversation, sharing answers or achieving similar benchmark scores is not enough to justify that specific label. Drawing on behavioural modelling, we propose five practical tests for deciding when claims of behavioural human-likeness have genuinely been earned.
Source: PsyArXiv (OSF Preprints)
Authors: Thomas Hancock, Jamal Amani Rad, Stephane Hess
Language: English
Subfield: Social Psychology
Field: Psychology
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7164773954
