safety-rules

This paper introduces BoardSafety, a benchmark built from vendor datasheets to evaluate whether LLMs can judge if an operation is safe for a specific development board. Results show current models, even with access to board-specific rules, cannot reliably make safe determinations.

The article discusses the challenge of hard-coding safety rules into AI models to prevent rogue actions, proposing simple rules as debate points.

The paper proposes GuardEn, an executable safeguard framework that uses safety-rule entailment to enhance visual compliance in vision-language models, achieving notable performance improvements.

Chemical accidents are rising as the Trump administration proposes weakening safety rules, drawing criticism from environmental groups and highlighting ongoing hazards.