Cited, and still re-checkable: AIScrapeSafe shows up in a governance whitepaper
June 24, 2026 · In the research
A research paper just cited AIScrapeSafe as a source. Not a screenshot pasted into a footnote, but a live verdict, referenced by the public License ID you can open and re-run for yourself.
The paper is Governed Autonomy: Human Accountability Above the Loop in Agentic AI, a 36-page whitepaper with 91 cited sources (Cougias, 2026; DOI 10.13140/RG.2.2.33030.74565). One of those 91 sources is us:
> AIScrapeSafe. (2026). Usage license report for linkedin.com (License ID 01KVVFCTFGKYDQRF954BJV23YH; verdict: restricted). Open Controls / MoxyWolf.
The record, not the screenshot
Click that License ID and you don't get our word for it. You get the record itself: linkedin.com, verdict restricted at 95% confidence, with the evidence still attached. The robots.txt the check read, the Disallow: / it matched, and the twenty AI agents the file blocks by name (GPTBot, ClaudeBot, CCBot, Google-Extended, and the rest).
That's the whole point of the thing. A citation usually freezes a claim at the moment it was made and starts going stale immediately. A Registered License Record doesn't. The paper cited an ID, the ID still resolves, and the evidence is still there to argue with. Anyone (a reviewer, a co-author, a regulator) can re-open the lookup a year from now and check whether the verdict still holds.
Why a governance paper reaches for a usage verdict
Governed Autonomy is about keeping a named human accountable for what an AI system does, and one of its two axes is the data. An ingress gate asks whether you had the right to feed the machine the input in the first place. That question is exactly what an AIScrapeSafe verdict answers. The linkedin.com record is the worked example, a real "are we allowed to use this?" resolved to a restricted, evidenced, re-checkable answer instead of a guess.
We'll be straight about it: this whitepaper comes from the same house as AIScrapeSafe, so we're not claiming an outside lab blessed us. What we are claiming is the mechanism, that a usage-rights call can be written into the literature as a stable, public, re-verifiable record instead of a stale screenshot. We'd like to see records cited that way everywhere, including the ones that turn out to disagree with us.
The one-time answer was never the product. This is the record doing the job we built it for.