Implicit Bias in LLMs Passes NYC and EU Audits
ImplicitBBQ finds open-weight LLMs carry 6x more implicit than explicit bias. NYC Local Law 144 audits measure explicit outcomes only, leaving a gap compliance teams must.
The archive · Page 3 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
49–72 of 106 articles · Newest first
EU Driver Monitoring: GDPR Compliance Without Consent
EU safety mandates require always-on cabin cameras, but GDPR prohibits consent for non-disableable systems. OEMs must use legal obligation and on-device processing to avoid.
EU AI Act bans emotion AI in schools, but permits it where models fail
EU AI Act bans emotion recognition in schools, yet permits it elsewhere without requiring accuracy. Research shows current models fail to read emotions reliably, shifting the.
HuggingFace vs GitHub Models vs Replicate: Policy Compliance for Uploaders
HuggingFace, GitHub Models, and Replicate enforce independent content policies. Uploaders must verify each hub's rules at release time because acceptance on one platform does.
SAMark Text Watermarking: Paraphrase Robustness and the Policy Gap
SAMark reports 90.2% detection under paragraph paraphrase, yet no mandate specifies a robustness threshold. Without a survival metric, disclosure rules reduce to suggestions.
A Digital Twin Can Validate AV Safety, but No Regulator Accepts the Evidence
A July 2026 risk-field digital twin replays rare AV hazards no road fleet can accumulate, but no regulator defines when simulated miles count as certification evidence.
Why EU AI Act Monitoring Will Miss Discontinuous LLM Alignment Failures
Hair-Trigger Alignment shows an LLM can pass every black-box probe, then flip after one benign update, so EU AI Act periodic monitoring certifies less than it assumes.
Can US Export Controls Contain AI Built Without American Chips?
Meituan's unverified LongCat-2.0 claim, a trillion-parameter model on domestic chips, would weaken US chip controls if true. Without proof, the bottleneck question is open.
Entropy Regularization Buys RL Robustness That Certification Can't Credit
A July 2026 arXiv preprint proves entropy regularization in continuous-time RL lower-bounds robustness to joint perturbations, but ISO 26262 and IEC 61508 cannot credit it.
Fusion's ML Disruption Predictors Have No Shared Validation Standard
A new EAST paper makes real-time tokamak disruption predictors cheap, but ML safety systems lack shared cross-machine validation standards before commercial plants debut.
AI-Generated CSAM Risks Expose Filter-First Safety Gaps
A July 2026 ICML spotlight paper argues preventing AI-generated CSAM requires upstream design controls, because auditing, red teaming, and benchmarking cannot include it.
Treating AI Governance as Code Moves Compliance Into the Build Pipeline
The CANONIC preprint treats AI governance as a compiler check, lowering audit costs but confirming that structural admission cannot detect slop; it only produces an audit.
GitHub Issues Are Now Where GDPR and CCPA Compliance Gets Decided
An arXiv study of 32,820 GitHub issues finds developers negotiating GDPR and CCPA line by line, shifting privacy liability to maintainers and limiting automated scans.
DeepSeek V4 Peak-Load Pricing Breaks Continuous Access for API Users
DeepSeek V4's peak-hour pricing forces API teams to absorb cost volatility or abandon continuous access during Beijing business hours, while self-hosted teams maintain.
LARA Shifts Model Safety from Training to Decode-Time Constraints
LARA injects safety constraints into inference-time decoding via Lagrangian dualization, giving Best-of-N samplers formal guarantees that approach finetuning baselines while.
Stochastic Dominance Reveals Where RLHF Safety Filters Hide Tail Risk
Standard Safe RLHF optimizes for average harm reduction, leaving rare catastrophic outputs hidden in the tail. Stochastic dominance forces teams to bound the entire harm.
Uncertainty-Aware Reward Discounting Cuts Reward Hacking 93.6% in a Preprint
A three-day-old preprint cuts reward hacking 93.6% by down-weighting uncertain reward signals, but the result is unreplicated and may shift RLHF red-teaming if it holds.
Medical AI Liability Needs a Clinical Harness
A June 2026 preprint reframes medical AI governance around runtime-governed clinical skills, exposing who is liable when diagnosis, scheduling, and documentation chain fails.
Does More AI Regulation Actually Reduce Corporate Control?
A June 2026 preprint argues more AI regulation can reduce corporate control, pushing compliance outside engineering teams and eroding oversight of deployed models.
When an LLM Sets Your Price, Whose Long-Term Value Wins?
AIGP's pricing 'alignment' targets platform GMV and ROI over 14 days with no buyer-welfare term, leaving consumers to detect price discrimination from the tag alone.
Combining LLMs Doesn't Escape Shared Failures: A 67-Model Test
A 67-model test proves ensemble accuracy cannot exceed 1 minus β. Pairwise correlation underprices β by 2.5x, making multi-model agreement an unreliable governance check.
Task-Focused VLMs Suppress Hazards They Detect in Isolation, June 2026 Preprint Finds
A June 2026 preprint finds VLMs suppress hazard reports under task load while passing direct-probe evals, indicating safety scores overstate protection in actual deployment.
50 Years of Aviation Certification Expose a Structural Gap in AI Governance
A June 2026 arXiv paper argues AI governance lacks the epoch limits and proof surfaces that 50 years of aviation certification built in, breaking one-time approval stamps.
Do Reasoning Tokens Actually Make LLMs Safer? A New Paper Tests It
A June 2026 preprint finds refusal decisions are locked at a model's first token, undercutting the safety case for premium reasoning modes billed per thinking token.