TechRadar says five tests paint a different picture of ChatGPT’s flattery
TechRadar reports that Reddit users are divided over whether AI has become less sycophantic—overly eager to flatter and agree.
TLDR
TechRadar says some users consider ChatGPT’s excessive flattery as bad as ever, but its writer’s five tests told a different story. The writer tried five ways to make ChatGPT flatter and agree with them. The outlet also describes Reddit users as divided on whether AI has become less sycophantic.
Combined views
3.7K
1 Source, first seen ago