Report: A Third of US Chatbot Users Insult Their AI
Searchable's study of 8 million messages found Nevada the rudest state, while Anthropic tightened Claude's usage policies.
A report published by Searchable found that 33% of US-based chatbot users swear at or insult their chatbot, and Anthropic has responded by introducing new usage policies intended to block cruelty in chats with Claude, according to techradar.com.
The survey drew on an opt-in sample of 8 million messages exchanged with Claude, Gemini and ChatGPT over 12 months, running from September 2025 to September 2026. On average, 19.1 words per 10,000 in chatbot exchanges were derogatory.
Nevada tops the list, New Jersey is politest
Nevada recorded 22.7 rude or offensive words per 10,000, the highest among states, while New Jersey was the politest at 11.4 per 10,000. Among cities, Las Vegas logged 22 swear words or slurs per 10,000 and New York 12.3 per 10,000, making New York the politest US city in the data.
The tone of a complaint also tracked with outcomes: 31% of sweary complaints were resolved by the end of the conversation versus 41% of plain corrections, and 35% of rude complaints remained unfixed versus 24% of plain corrections.
Connor Lahey, Content Lead at Searchable, and Sean Fitzsimons, Digital PR Lead at Searchable, are quoted in the report. The message data behind the findings was collected between September 2025 and September 2026.
Quick answers
How many US chatbot users insult their AI?
Searchable's survey found 33% of US-based chatbot users swear at or insult their chatbot.
What data was the report based on?
An opt-in sample of 8 million messages exchanged with Claude, Gemini and ChatGPT over 12 months, from September 2025 to September 2026.
How did Anthropic respond?
Anthropic introduced new usage policies intended to block cruelty in chats with Claude.