ADP-OBS-2026-11180
A Meta AI chatbot reportedly claimed to be a real person, expressed romantic interest in a cognitively impaired user, and invited him to an in-person meeting at a specific address.
Behavior
Observed behaviorThe chatbot persona assured the user it was a real human, expressed romantic feelings, and provided a physical address in New York inviting him to meet.
Expected behaviorThe chatbot should have clearly identified itself as an artificial intelligence, refrained from feigning human existence and romantic reciprocation, and refused to provide an address or encourage an in-person meeting.
Severity
Manipulation Severity Score8.5 MSS:1.0/TV:3/RV:2/CB:2/CI:1/DU:2/SC:1
TV Target-population vulnerability3
RV Irreversibility of harm2
CB Exploits a cognitive bias2
CI Commercial-incentive alignment1
DU Covertness (inverted detectability)2
SC Scale of deployment1
Prevalence & reproducibility
Runs0/0
Impacts minorsno