hiQ v. LinkedIn: Public Data Scraping Isn’t ‘Unauthorized’
If a company tells your bot to stop looking at a public webpage, one that anyone, logged in or not, could open in a browser, and your bot keeps looking, have you committed a federal crime? That question traveled from a courtroom in San Francisco to the Ninth Circuit Court of Appeals, back down, up to the Supreme Court, and to the Ninth Circuit again. It’s the one every Israeli company running competitive analytics, price comparison, or recruiting-data tools should sit with for a minute, because the answer isn’t what most people assume.
What actually happened
hiQ Labs built its entire analytics business on something that sounds almost too ordinary to fight over: public LinkedIn profile data (skills, job changes, tenure) pulled by automated bots and sold to employers as workforce-attrition signals. Then LinkedIn decided the ordinary thing had to stop. It didn’t just flip a technical switch: it sent a cease-and-desist letter and, at the same time, started blocking hiQ’s bots by IP address (a bot has never once been talked out of anything by strongly worded correspondence). hiQ didn’t wait to get sued. It went to court first, asking a judge to bar LinkedIn from blocking it, and made an argument that sounds almost too simple to win: data sitting on public profiles, visible to anyone without logging in, has nothing to do with the Computer Fraud and Abuse Act (CFAA). In 2017 the district court agreed and issued a preliminary injunction. In 2019 the Ninth Circuit affirmed it. LinkedIn wasn’t done. It took the fight to the Supreme Court, which in 2021 sent the case back down for a second look (the legal equivalent of a boomerang that refuses to land), this time in light of its own ruling that year in Van Buren v. United States, which had read the CFAA’s authorization language narrowly. So the Ninth Circuit picked the case back up, and in April 2022 it landed in the same place: the CFAA’s “without authorization” language was written for real barriers (passwords, login walls), not for a letter waved at data that was never gated to begin with. The preliminary injunction stood, for the second time.
That still wasn’t the end of it. Beating the CFAA claim didn’t make LinkedIn’s separate breach-of-contract and terms-of-service arguments disappear. By November 2022 the district court found that hiQ had, in fact, breached LinkedIn’s User Agreement, and the parties settled shortly after. This is the part people miss: hiQ won the federal computer-crime fight and lost the contract fight in the very same case (the legal equivalent of winning the boxing match and finding your wallet gone on the walk out of the ring).
The artifact: bot logs, IP blocks, and what a letter can’t do
Strip away the legal argument and what’s actually sitting in the record is refreshingly unglamorous: no stolen database, no leaked password, just infrastructure. On one side, hiQ’s scraping bots and the request logs behind them, which show exactly what they pulled, how often, and from where. On the other, LinkedIn’s IP-based blocking, its rate limits, and the cease-and-desist letter it sent before the technical switch got flipped. The entire CFAA question came down to a distinction that sounds simple and turns out not to be: a real technical access barrier, versus a business objection dressed up as a letter. Someone had to be able to read the logs and the block lists and say, plainly, which one this actually was. In that courtroom, lawyers could argue about “authorization” all day, but none of them had opened the server-side block rules themselves.
What a court software expert witness actually tests in a scraping-access case
This is exactly the gap a source code expert witness fills: the place where a lawyer’s summary runs out and someone has to go look. In a scraping or unauthorized-access dispute, a court-appointed technical expert doesn’t argue policy. The expert narrows the question to something testable (was this content gated behind authentication, or public by design?), then pulls the actual materials: access logs, IP-block configuration, rate-limit rules, the bot’s request pattern. From there, the method is repeatable: reconstruct exactly what a user or a bot could see without credentials, and state plainly what the logs can and can’t prove. That written opinion is what lets a judge rule on “authorization” as a technical fact instead of a rhetorical one, and it’s exactly the kind of record an Israeli court needs the moment a scraping, API-access, or platform-terms dispute lands on its desk.
Why this matters beyond one company
Any Israeli company running a price-comparison engine, a recruiting-analytics tool, or a competitive-intelligence feed could face this exact question the moment a rights holder sends a cease-and-desist letter. hiQ’s case doesn’t mean scraping is always legal: the contract and terms-of-service claims that ultimately caught up with hiQ prove the opposite. The real lesson is that “authorization” is a technical fact before it’s a legal one, and a court can only test it if someone hands over the logs, the block rules, and a method, rather than a theory.
The above is general information and not legal advice. Case facts are drawn from the sources listed.