# NYC's idea of fining chatbot 'doom loops' should be the default everywhere: every support bot needs a guaranteed human exit within 3 messages — retained record

> Retained reddit record from reddit:artificial.

Canonical: https://highsignal.app/data/records/44009318a0923f2f

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:artificial)  published 2026-11-17 14:16 ingested 2026-11-18 02:36

[original source ↗](https://www.reddit.com/r/artificial/comments/1xyljcr/nycs_idea_of_fining_chatbot_doom_loops_should_be/)

## retained text



```
Last month I spent 22 minutes in a bank's support bot trying to report a card I'd lost. Every path led back to 'Would you like to check your balance?'

Bots are great for 'where's my order'. They're dangerous for anything urgent, financial or medical. A simple rule (say 'human', or fail to resolve twice, and you get escalated) costs companies little and fixes the worst cases.

Would you support a rule like that? Or does mandating humans just push companies to hide the bot better?
```



document fetched 2026-11-18 00:20 · canonical https://reddit.com/r/artificial/comments/1xyljcr/nycs_idea_of_fining_chatbot_doom_loops_should_be
