New research suggests being rude to AI makes it less likely to fix your problems
Reported by TechRadar and Android Headlines2 sources ↓
Searchable analyzed more than 8 million messages and found chatbots resolved a complaint in the next reply 24% of the time when it included insults or swearing, versus 35% without them.
Key points
- Searchable found one in three American chatbot users had sworn at or insulted a chatbot
- The most common triggers were a chatbot not following instructions (19%), giving a wrong answer (18%) or misunderstanding a request (15%)
- Men used a third more swear words and insults than women, and users aged 35 to 44 swore most often
Sources
Details
Searchable’s analysis covered messages sent to AI tools over one year. The company reported that complaints without insults or swearing were resolved in the next reply 35% of the time, compared with 24% when the complaint included them. The research does not establish why the response rates differed.
The findings come as Anthropic’s updated usage policy for Claude took effect on November 12. Anthropic says it applies to repeated, extreme abusive behavior without a discernible purpose, not ordinary frustration, pushback, creative themes or model testing; the company says Claude may end conversations with persistently abusive users.
Written from coverage as of
Images
- TechRadar
- TechRadar
Stocks in this story
Change since open: 20 Nov to 21 Nov 2026
Last refresh