Google Keeps Fighting AI Web Scrapers After Court Loss. Here's Why It Matters.
Google sued a company called SerpApi for selling search results it had scraped from Google. A judge dealt Google a setback last week. Google says it is pressing on anyway, and Reddit has quietly joined the fight.
Key points
- Google filed a lawsuit against SerpApi in December 2023, accusing the company of scraping and reselling its search results.
- Google invoked the DMCA, the Digital Millennium Copyright Act, a law usually used to protect music and film from piracy, to argue SerpApi broke through its anti-scraping defences.
- A court dealt Google a significant blow last week, but Google confirmed it will not drop the case.
- Reddit is involved as a named party with an interest in blocking unauthorised scraping of its content.
- The outcome could affect how any AI company collects the data it needs to train its systems.
Google took a hit in court last week. A judge ruled against it in a lawsuit it had filed against a company called SerpApi, which sells a service that pulls data from Google search results and packages it for businesses to use. Google says that is theft. SerpApi says it is just reading publicly available web pages.
Despite the loss, Google confirmed it will keep fighting.
What exactly did SerpApi do?
SerpApi built a product it marketed as a "Google Search API", a piece of software that lets paying customers query Google search results automatically, at scale, without going through Google directly. The problem, Google says, is that SerpApi broke through the technical barriers Google had built to stop exactly that kind of automated harvesting.
Google filed the suit under the DMCA, the Digital Millennium Copyright Act, a law passed in 1998 that makes it illegal to bypass copy-protection systems. The law is most often used against people who crack DVD encryption or pirate streaming services. Google argued its anti-scraping technology counts as that kind of protection.
The content at stake is not just blue links. Google highlighted its "knowledge panels", the information boxes that appear when you search for a well-known person or company. Some of that information is licensed from third-party rights holders who pay to have their data appear in Google results. Google argued that SerpApi's scraping threatened those licensing deals.
Reddit, which also licenses its content and has its own scraping concerns, joined the case as an interested party.
Why does a copyright law designed for DVDs apply here?
That is exactly what the court questioned. Using the DMCA against a web scraper is an unusual legal move, and the judge's scepticism last week reflects how untested this theory is. Google is essentially asking a court to agree that anti-bot software counts as a copy-protection system the same way DVD encryption does. No court has firmly said yes to that yet.
If Google eventually wins that argument, it would give every major website a powerful new tool to block AI companies from training on their content without permission.
What does this mean for you?
For most people, nothing changes today. But the case sits at the centre of a fight that will shape how AI tools are built tomorrow. AI models, the software engines behind tools like ChatGPT and Google's own Gemini, learn by processing enormous amounts of text scraped from the web. If courts decide that scraping search results, or eventually any website, breaks copyright law, the cost of building those tools goes up sharply.
Higher costs for AI companies tend to flow downstream: slower product updates, pricier subscriptions, or simply fewer players left standing.
As Ars Technica first reported, this case is one of several tests playing out right now over who owns the data that trains AI.
The honest takeaway: if you run a website or create content professionally, watch this case. It is slowly building the legal framework that will decide whether AI companies have to pay for the data they use.
