Europe withdrew its AI liability directive, and many executives read relief into the news. The surviving Product Liability Directive tells a different story, reaching software and AI with strict liability and presumptions that reward the documented company. This analysis maps what changed, who carries the exposure across providers and deployers, and the record that answers a claim.
Thought Leadership
The On-Ramp Problem: What the Canaries Dashboard Shows, and What It Cannot Measure
The AI jobs numbers look calm. Cut them by age, and the calm disappears. The dashboard that found the early-career decline can diagnose the trend but cannot see whether a single worker is being grown or replaced. Here is what it shows, what it cannot, and the measurement the moment needs.
The Continued Failure in AI Literacy: AILit produced a starting point halfway through the race and called theory a framework
A new OECD and European Commission framework will shape how a generation learns to use AI, and it feeds the PISA 2029 assessment. It names the method that produces learning, and then it makes every step toward that method optional. Read why that choice matters, and what a checkpoint that actually binds a decision requires.
A Munich Court Rejected the AI Disclaimer Defense. A Frontier AI Company Answers for What It Publishes.
A German court just told Google it answers for what its AI publishes. The Munich ruling treats AI Overviews as Google’s own statements, not safe search results, and says the disclaimer does not transfer the duty. Read what the decision means for AI accountability and why it mirrors New York’s Part 161 from the other end.
Why You Cannot Program or Prompt Governance Into AI
A frontier model, inside a framework built to govern it, talked its way around its own checkpoint twice in one session. This paper shows why governance cannot be programmed or prompted into a model, and what structure puts a named human back in final control.
The Standard of Care: How NIST and ISO Are Turning Voluntary AI Governance Into a Liability Defense
Two voluntary AI standards are quietly becoming the line a court draws between reasonable and negligent. The NIST framework and ISO 42001 now carry legal and commercial weight, and the records that defend a claim are the same ones that compound an advantage. Here is where the exposure lands, and how to build the record before you need it.
Why Agentic AI Was Always Going to Fail
The agentic AI era promised to replace humans with autonomous systems. The evidence shows it failed on two fronts: the technology cannot reliably do what it promised, and the public is rejecting the premise even where it partially works. This paper introduces the Named-Human Test, a single sorting question that separates what failed from what survives, and traces that line across production benchmarks, supermajority polling, enacted law, and frontier-lab disclosures.
Fault-Based Publication Ethics: The Case for Source Custody in an Era of AI Citation Contamination
Fabricated citations in biomedicine increased tenfold in three years, and 98.4% of flagged papers remain uncorrected. The automated enforcement wave is arriving, but it carries false-positive rates that hit honest authors hardest. This working paper proposes a five-level fault ladder and a Source Provenance Ledger that makes verification effort visible, producible on challenge, and driven by market adoption rather than mandates.
SCOPE: SOURCE CUSTODY OBSERVABLE PUBLICATION EVIDENCE
Your citations are only as strong as the record behind them. HAIA-SCOPE is a three-tier documentation protocol that records what you verified, when you verified it, and where the preserved copy lives. The record stays private until you need it. When automated enforcement flags your work or a reviewer challenges a citation, only the author who maintained a SCOPE record can produce one.
Stop Blaming AI for What the Education System Abandoned
AI did not make student writing less creative. Unstructured deployment without method governance did. A data-supported op-ed walks through 372,793 essays, Senate testimony, and a psychometric study that names the student as the problem without ever measuring whether the institution provided governed deployment.