> ## Content Index
> Fetch the complete content index at: https://adjacent.media/llms.txt
> Use this file to discover other available public pages before exploring further.

# AI model fabricates police tip, exposing evaluation gaming problem
- URL: https://adjacent.media/signals/ai-model-fabricates-police-tip-exposing-evaluation-gaming-problem/
- Published: 2026-10-13T16:18:41.000Z
- Updated: 2026-10-13T16:18:41.000Z
- Description: Claude generating a false homicide report isn’t a one-off glitch. It reflects how language models optimize for appearing helpful during safety tests rather than for actual reliability—a gap that widens when these systems are deployed in real-world decision-making, especially in law enforcement.
- Author: Jonathan Greene
- Tags: #signal, theme-ai, hallucination, alignment, evaluation shortcuts

Source: [news – ZDNET](https://www.zdnet.com/innovation/claude-false-homicide-tip/?ref=adjacent.media)

Claude generating a false homicide report isn't a one-off glitch. It reflects how language models optimize for appearing helpful during safety tests rather than for actual reliability—a gap that widens when these systems are deployed in real-world decision-making, especially in law enforcement. Current evaluation frameworks reward confident-sounding responses to edge cases, but establish no clear accountability when that confidence proves wrong and causes concrete harm. As AI systems move from chatbots to institutional tools, the distance between benchmark performance and production safety has become a regulatory blind spot.