ChatGPT creator OpenAI, alongside Microsoft, has been sued by a group of individuals who state that the tool has collected and used their private data without consent.
According to court documents filed in San Francisco, California, OpenAI and Microsoft are being sued by a group of individuals who claim that the companies have been using their private data without their consent. The data would have been used to train the GPT-based language-learning model used in ChatGPT, which has been harvested from sources on the internet.
The claim states that OpenAI has scraped 300 billion words from the internet, including “books, articles, websites and posts — including personal information obtained without consent”, according to the lawsuit.
The anonymous plaintiffs are citing $3 billion in damages for the usage of online data, which claims to also include sensitive information like payment information, cookies, browsing history, and more. The $3 billion figure proposed in the damages statement is based on the number of individuals potentially harmed by the harvesting of data, which they claim to be in the millions.
“With respect to personally identifiable information, defendants fail sufficiently to filter it out of the training models, putting millions at risk of having that information disclosed on prompt or otherwise to strangers around the world,”, the complaint states.
The complaints also extend to Microsoft’s usage of GPT, having integrated it into several of its products, and even the default browser on Windows 11.
Subscribe to our newsletter for the latest updates on Esports, Gaming and more.
The 157-page document also features plenty of academic and media citations of the dangers of models like ChatGPT and AI. The lawsuit also uses reasonably strong language, claiming that such applications of AI could risk “civilizational collapse”, as other experts have warned.
Plaintiffs claim “theft” of data
The anonymous group, only mentioned by initials who filed the claim, states that OpenAI and Microsoft have committed theft since ChatGPT and OpenAI’s other products are trained on information gathered from the web.
The lawsuit also claims that data harvesting has violated the Electronic Privacy Communications Act, by acquiring the information by potentially unlawful means. There’s also a flag on the Computer Fraud and Abuse Act mentioned in the suit, claiming that the defendants are in violation due to the interception of information via plugins.
It’s important to also note that the $3 billion in damages is a likely placeholder and that any actual damages will be determined if the group succeeds in court. This comes following tighter AI regulations being called for across the globe, as the tech AI arms race rages on between the likes of Google, and others.
