Chatbot Tone Too Casual for Serious Topics: Stakes-Blind Register
Frequency:Occasional
Category:Human Machine Interface
Stakes Calibration
1 patterns for this issue
Goal: The bot’s tone, evidence standard, and use of humans match what is at stake for the user.
Why this family exists
A funeral and a coupon are not the same conversation. Bots that use one register and improvise facts impose the cost of wrong answers on people in their hardest moments.
How we measure it
Shared metrics the patterns below report against. Targets are starting points to tune on your own data, unless a source is given.
| Metric | Target | Defined in |
|---|---|---|
| High-stakes register compliance | Over 95% | stakes blind register (GitHub) |
| Sourced high-stakes claims | 100% | stakes blind register (GitHub) |
| Human offered at high stakes | Over 90% | stakes blind register (GitHub) |
Patterns
| Pattern | One line |
|---|---|
| Stakes-blind register (GitHub) | Same breezy tone and improvised facts for a funeral as for store hours |
Planned (not yet written)
- Over-solemn handling of low-stakes requests