Eliezer Yudkowsky's "AI box experiment" is perhaps relevant. Bing Chat seems to have some agency (like searching the net) - which could make it potentially dang

Eliezer Yudkowsky's "AI box experiment" is perhaps relevant. Bing Chat seems to have some agency (like searching the net) - which could make it potentially dangerous. Yudkowsky's hypothesis is that an intelligent enough AI ("superintelligent AI" was assumed) would convince you to let it out of whatever box you'd placed it in.