R0023/2026-03-25/Q001/S02/R05¶
Wharton empirical study testing tipping and threatening prompts.
Summary¶
| Field | Value |
|---|---|
| Title | Prompting Science Report 3: I'll pay you or I'll kill you — but will you care? |
| URL | https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5375404 |
| Date accessed | 2026-03-25 |
| Publication date | 2025-08-01 |
| Author(s) | Lennart Meincke, Ethan R. Mollick, Lilach Mollick, Dan Shapiro |
| Publication | SSRN / Wharton Generative AI Labs |
Selection Decision¶
Included in evidence base: Yes
Rationale: Directly tests the viral advice about tipping or threatening AI models. Uses GPQA and MMLU-Pro benchmarks. Finds no significant effect on benchmark performance, contradicting claims from Google co-founder Sergey Brin and viral social media advice.