Showing cs.CLShow all
3 papers · 1 filter
cs.CL2026
The inattentional gap in task conditioned AI models that omit otherwise reportable safety critical signals
Kwan Soo Shin
AI in radiology and other safety-critical workflows is evaluated on the hazards it is told to find, yet harm arises disproportionately from hazards no one specified. We show that c…
cs.CL2026
The Reasoning Trap: An Information-Theoretic Bound on Closed-System Multi-Step LLM Reasoning
Kwan Soo Shin
When copies of the same language model are prompted to debate, they produce diverse phrasings of one perspective rather than diverse perspectives. Multi-agent debate (MAD), and mor…
cs.CL2026
The Compliance Gap: Why AI Systems Promise to Follow Process Instructions but Don't
Kwan Soo Shin
An auditor instructs an AI assistant: "open each file individually using the Read tool -- no scripts, no agents." The AI replies "Yes" -- then issues a single batched call summariz…