Microsoft and OpenAI accused of stealing user identities for AI training
Companies ignored legitimate ways to buy AI data and decided to collect it for free, plaintiffs say.
Microsoft And Open AI were at the center of a legal scandal over their artificial intelligence products, which 16 anonymous plaintiffs allege were illegally using and distributing their personal data without proper notice or consent.
IN lawsuit the companies are said to have ignored legitimate ways to obtain data for their AI models and chose to collect it for free. According to Plaintiffs, “Despite established protocols for the purchase and use of personal information, Defendants took a different approach – theft.”
Companies systematically extracted 300 billion words from the Internet, books, articles, websites and messages, including personal information obtained without consent. OpenAI did this in secret and without registering as a data broker, as required by applicable law.
The plaintiffs allege that through their AI products, Microsoft and OpenAI collect, store, track, share and disclose the personal information of millions of people, including product details, account information, names, contact details, login credentials, emails, billing information, transaction history, browser data, social media information, chat logs, usage data, analytics, cookies, search queries and other online activity.
The lawsuit alleges that Microsoft and OpenAI also embedded user information in their AI products that reflects hobbies, religious beliefs, political opinions, voting results, social media and support group memberships, sexual orientation and gender identity, employment records, family photos, friends. and other data resulting from online interactions.
Plaintiffs allege that Defendants fail to sufficiently filter personally identifiable information from learning models, exposing millions of people to the risk of disclosing that information to strangers around the world.
The complaint alleges that Microsoft and OpenAI violated:
- US Electronic Confidential Communications Act, obtaining and using personal information, and illegally intercepting communications between users and third-party services through integration with ChatGPT and similar products;
- Computer Fraud and Abuse Act by intercepting interaction data using plug-ins;
- California Invasion of Privacy Act; Unfair Competition Act;
- Illinois Biometric Privacy Act;
- Consumer Fraud and Misleading Business Practices Act;
- New York Business Law;
- negligence law;
- Law on unjust enrichment.
The complaint is filled with quotes from the media and academic documents expressing dismay at AI models and ethics, but the lawsuit makes little mention of specific cases of harm. The plaintiffs are seeking damages in the amount of $3 billion, although this figure appears to be provisional. Any actual damages will be determined if the plaintiffs win based on the court’s findings. Microsoft and OpenAI declined to comment.
Recently details of the case against Microsoft, GitHub and OpenAI appeared about the documented addiction of a programming assistant GitHub Copilot reproduce public code protected by copyright. Microsoft’s GitHub platform, owned by Microsoft, reportedly deliberately configured Copilot so that the tool generates small changes to the code provided to developers, and so that the output is not marked as a direct copy of existing software.
In addition, OpenAI faced a libel suit from radio host Mark Walters, who accused it of chatbot ChatGPT spread false information about his participation in the criminal case. Walters filed a lawsuit in a Georgia state court seeking compensation for moral damages and removal of false information.