The Right to Scrap Data on the Internet: From the US Case hiQLabs, Inc. v. LinkedIn Corp. to the ChatGPT Scraping Cases: Differences Between US and EU Law

Abstract

Data scraping is of the essence for generative artificial intelligence (AI), such as ChatGPT, as the data needed for AI training are in most cases obtained by this means. The first litigations regarding data scraping for training of generative AI systems are starting now. It is therefore relevant reviewing the case Law and regulation regarding data scraping, including the United States (US) Court of Appeals of the Ninth Circuit in the hiQ v. LinkedIn case. This case reveals the important differences between the US Law and the European Union (EU) (and United Kingdom (UK)) Law under several points of view: criminal law, privacy/data protection law, intellectual property (IP) rights (and sui generis rights), and even law of contracts. The idea that there is no ownership to unprotected, non-personal, publicly available data coincides with some recent EU Member States rulings. Data scraping is not illicit as such (the Google search engine is always scraping, and has been doing so since its creation), but it needs to respect some criminal law restrictions, privacy or data protection obligations, copyright and contractual Law. This article reviews this protection in both the US and the EU/UK and analyses the impact of the existing data scraping case law in the ChatGPT issues and very recent litigations. Data Scraping, AI, ChatGPT, hiQ v. LinkedIn, Data Protection, GDPR, Intellectual Property