WikiHow Sues OpenAI Over Unauthorized Scraping of 11,000 Tutorial Articles
WikiHow has filed a lawsuit against OpenAI in a Manhattan federal court, alleging the unauthorized scraping of over 11,000 tutorial articles to train ChatGPT and GPT models. The lawsuit claims that this scraping infringed on at least 1,200 of WikiHow's registered copyrights. This lawsuit adds to the growing wave of copyright litigation against AI companies regarding the use of web-scraped training data. It highlights the ongoing conflict between AI developers relying on fair use and content platforms defending their intellectual property and business models. WikiHow argues that OpenAI's models can now generate competing tutorial content at a fraction of the time and cost, which directly reduces WikiHow's revenue. OpenAI has responded by stating that its models are trained on publicly available data and that their training practices constitute fair use.
## BACKGROUND
WikiHow is one of the world's largest collaborative "how-to" guide websites, containing step-by-step instructions on a wide range of topics. Generative AI models require massive amounts of text data to learn language patterns, which developers often gather by scraping websites, leading to legal challenges over whether this practice violates copyright laws.