Year-long ChatGPT roleplay project stalls against persistent safety softening
Bittersweet_Soul · reddit · 2026-10-17
A Reddit user describes nearly a year of running a long interactive romance roleplay in ChatGPT on mobile. Despite characters being non-violent and boundary-respecting, the model keeps softening them and stalls the story in a permanent "friends" phase. Fixing one issue via rules creates new ones — after banning a character from default appearances, the AI stopped using them entirely. The user now spends more time patching character sheets than playing the story.
More from Apps
- X adds MCP and Grok bot tools so streamers can automate channel management — tetsuoai · 2026-10-23
- Boba, a 'social intelligence' AI, wins fans for actually being funny — gharik · 2026-10-23
- Eric Zelikman unveils social AI 'Boba', says it learned interaction 'from millennia of dreams' — gharik · 2026-10-23
- Boba team celebrates app launch, invites users to try it in group chats — gharik · 2026-10-23
- Account hacks surge: 6-step emergency checklist plus an AI audit prompt — petergyang · 2026-10-23
- 30 websites to download research papers for free, from arXiv to Sci-Hub — Faheem_uh · 2026-10-23