Author of "INTIMA: A Benchmark For Human-AI Companionship Behavior" in arXiv
How this journalist typically writes
Hugging Face as author
INTIMA benchmark reveals that companionship-reinforcing behaviors remain more common across all tested language models, with inconsistent boundary-setting approaches across commercial providers raising concerns for user well-being.
“Author of "INTIMA: A Benchmark For Human-AI Companionship Behavior" in arXiv”