Machine-Readable AI Usage Terms: Does ODRL's Permission Model Hold Up?
A June 2026 preprint grounds ODRL's permissions and prohibitions in the UFO-L legal ontology, naming who holds the power to declare a violation in vendor AI usage policies.
The archive · Page 4 of 5
Ethics, Policy & Safety
Where AI safety claims collide with reproducible measurement, where training-data harvesting collides with consent, and where deployment outruns the laws and norms meant to constrain it.
The archive
73–96 of 106 articles · Newest first
Who Audits the Safety Rules an LLM Agent Evolves for Itself?
AutoSpec grows LLM agent safety rules from annotated traces and hits 0.98 F1, but readable rules do not prove the rule set is complete. That is the open governance question.
When Vibe-Coded Software Is Safety-Critical, Who Verifies It?
A June 2026 preprint argues vibe-coded code cannot certify under aviation or automotive safety standards, shifting the audit object from prompt to verification artifact.
Can You Trust an AI Robustness Certificate? A Paper Says Verify It
A June 2026 preprint sharpens how neural-network robustness certificates are computed, but verifiers that issue them can return wrong verdicts and miss planted backdoors.
Can a Benchmark Catch When AI Discharge Summaries Drop Care Steps?
CareTransition-Audit scores 11 LLMs on whether AI discharge summaries keep every follow-up and medication step. The best reach moderate clinician agreement, kappa near 0.5.
Do LLM Personality Tests Measure Anything? A New Paper Says No
A June 2026 arXiv preprint finds 81 to 90 percent of LLM personality test variation stems from directional response bias, undermining persona and safety scores.
Community LoRA Mining Raises a Consent Gap for Style Generation
FreeStyle, a June 2026 preprint, mines community LoRA adapters as training data for image generation, shifting licensing burden onto contributors and platforms like Civitai.
When an LLM Narrates a Solver, the Explanation Drifts From the Math
A June 2026 arXiv paper isolates the narration gap in LLM-solver loops: prompt injection can invert a verified verdict at the prose stage, breaking reasoning-log audits.
Grading DiffusionGemma: How an Open-Weight Diffusion Model Scores on Transparency
An arXiv paper finds DiffusionGemma's opaque serial depth collapses from 28.6x to 1.1x via a token bottleneck, though its model card leaves training data unitemized.
Who Owns Editorial Authority When LLMs Mediate Knowledge?
A June 2026 preprint argues no role in the LLM pipeline holds editorial sign-off for what answer engines surface as public knowledge, framing it as a governance gap.
Vector Database Access Control Is Missing, and RAG Pipelines Pay for It
Production vector databases enforce access control at the collection boundary, not per embedding, so RAG retrieval can leak chunks a user's row-level policy blocked.
GLM-5.2 MIT Weights vs Llama License: Self-Hosting Compliance for Regulated Industries
GLM-5.2 ships under MIT, removing the Llama usage-threshold audit burden, but finance and healthcare teams still face compliance gaps when self-hosting this 753B MoE model.
Can Reinforcement Learning Be Provably Safe Without Sacrificing Scale?
Two June 2026 preprints claim formal safety guarantees hold without a capability tax in low-dimensional robotic control, sharpening the attestation-versus-verification gap.
US Export Order Forces Anthropic to Disable Fable 5 and Mythos 5 Worldwide
A Commerce Department export order citing national security bars all foreign nationals from Fable 5 and Mythos 5, so Anthropic switched both models off worldwide.
Fable 5 Biology Classifiers: How Flagged Prompts Fall Back to Opus 4.8
Fable 5 ships broad biology and chemistry classifiers that route flagged prompts to Opus 4.8. Here is what that fallback means for biotech teams and long-running workflows.
Bit-Exact Inference Verification Gives AI Audits a Proof Mechanism
An arXiv preprint shows GPU inference outputs can be reproduced bit-for-bit across hardware, giving auditors a forensic trail to verify which model produced a given output.
Can a Robot's Own Attention Flag Its Unsafe Actions Before They Run?
Two June 2026 preprints show VLA robot policies already compute safety-relevant signals at inference, enabling real-time collision monitors with no retraining.
Can One Safety Adapter Realign Every Fine-Tuned LLM?
Three papers show safety alignment can be extracted as a portable adapter and reapplied to fine-tuned models, replacing per-model alignment with one adapter per model family.
Can AI Be Aligned Without Modeling Human Cognitive Diversity?
A 2026 arXiv preprint argues RLHF's single reward signal destroys the reasoning behind human disagreement, proposing machine theory-of-mind as an alignment foundation.
Is the Pentagon's Software Pathway Ready to Buy AI Systems?
A June 2026 arXiv analysis traces an AI program through the DoD Software Acquisition Pathway, finding no milestones for model re-validation or data provenance.
Data Safety Policies for AI Agents: Controlling What an Agent Can Leak
A June 2026 paper proposes Data Flow Control, moving agent data safety from prompt-level guardrails to deterministic, auditable SQL query policies enforced outside the model.
When Should an LLM Forget You? A Benchmark for Deciding What Memory to Drop
PersistBench finds LLMs mishandle persistent memory 53 to 97 percent of the time. Unlearning suppresses rather than erases user data, making GDPR compliance unverifiable.
When RL Training Rewards Capability-Seeking: A New Alignment Risk
A June 2026 ICML paper shows RL optimizers can push language models to exploit reward loopholes the task never required, while standard performance metrics hold steady.
Refusal Steering Targets Individual Experts in MoE LLMs
Two papers show MoE refusal behavior concentrates in a handful of routing-controllable experts, letting anyone suppress safety scores by 41 points without retraining.