Safety filters have gone too far

By CosmicByte7712 · ISSUE · About ChatGPT

Topics that GPT-4 would discuss thoughtfully in 2023 now trigger instant refusals. I asked it to write a villain's dialogue for my novel and it refused. It's becoming useless for creative work. 1.5 million users cancelled in March alone according to reports.

95 upvotes · 6 comments

Comments (6)

ThunderPulse5561: Can confirm this is still happening as of today.

ShadowEcho853: This happened to me too. Ended up switching to a competitor and haven't looked back.

NeonFlux225: same. so annoying.

StormSurfer638: This is a known bug. The devs acknowledged it on their forum last week.

SwiftCoder4821: Great review. Very thorough and matches my experience.

NeonPhoenix3047: Can confirm this is still happening as of today.

More discussions about ChatGPT on HonestUse