Google has filed a lawsuit against SerpApi, a company that offers an API allowing users to simulate human search queries. This move highlights ongoing tensions over access to data used for training and running AI models. Many large language models are built on web-scraped data, often without the permission of website owners, leading to legal battles over copyright and data rights.

Google’s Dual Role in Data and AI Development

Google is in a unique position because it both collects and curates massive amounts of search data, and also develops its own AI models like Gemini. These models are integrated into various Google services, including search. While Google’s search results mainly come from scraping websites itself, it also licenses and creates content that it uses within its platform.

However, other companies and AI developers want access to that same data to build competing products. Google views this as a threat, especially when third-party firms like SerpApi are involved in scraping content that Google has licensed or generated. The lawsuit claims that SerpApi bypasses security measures to access protected content and resell it for profit.

What Google Accuses SerpApi Of Doing

In a blog post, Google’s General Counsel Halimah DeLaine Prado explained that SerpApi allegedly “circumvent[s] security measures” that protect copyrighted content appearing in Google search results. She added that the company is seeking to stop SerpApi’s bots, which scrape content without respecting the rights or preferences of website owners and content providers.

Google states that while it mostly gathers search data itself, the lawsuit targets SerpApi’s use of licensed content like images in Knowledge Panels and real-time search features. Google argues that SerpApi resells this data for a fee, ignoring the rights of the original content creators and webmasters. The company emphasizes that this practice violates the rights of content owners and Google’s licensing agreements.

SerpApi’s Response and Broader Implications

SerpApi denies any wrongdoing, saying it provides public search data that anyone can access through a regular browser. The company claims that its services simply give developers, researchers, and businesses access to publicly available information. They argue that their activities are protected by the First Amendment and that they follow all relevant laws, including fair use principles.

SerpApi also states that their business model is legitimate and plans to defend itself vigorously in court. They believe the lawsuit attempts to stifle innovation by companies that rely on their data to develop next-generation AI and other applications. This legal action could mark a turning point, signaling that AI firms may soon face more restrictions and legal challenges when sourcing web data.

There is particular concern about how SerpApi’s clients, like OpenAI and Perplexity, benefit from this data access. Reports suggest that some of these AI companies are already using SerpApi’s services to improve their language models. As legal battles continue, the future of how AI companies access and use web data remains uncertain, with potential impacts on innovation and competition in the tech industry.

Inspired by

• https://www.computerworld.com/article/4111155/google-says-no-to-training-ai-on-its-search-results.html

Sources

• shutterstock.com
• infoworld.com
• techcrunch.com
• pcworld.com
• blog.google