OpenAI 'unprecedented web scraping' trains AI models


Can you improve this page?
Share your insights with us

A lawsuit filed against OpenAI in California, USA, alleged that two of its AI models, ChatGPT and DALL-E, were trained using hundreds of millions of people’s data without proper consent. 

The160-page complaint, served on behalf of 16 plaintiffs, accused OpenAI of training its generative AI programmes ChatGPT and DALL-E on 'stolen private information' taken from hundreds of millions of internet users, including children, without proper permission.

The lawsuit argued that OpenAI integrated its systems with third-party platforms like Snapchat, Spotify, Stripe, Slack, and Microsoft Teams, enabling OpenAI to secretly gather users’ images, locations, music tastes, financial details, and private communications. 

The suit also argued that this data collection violated the terms of service of these platforms and privacy laws and constituted unauthorised access to people’s information.


Developer: OpenAI
Country: USA
Sector: Media/entertainment/sports/arts
Purpose: Generate text; Generate images
Technology: Chatbot; NLP/text analysis; Neural network; Deep learning; Machine learning; Reinforcement learning
Issue: Privacy
Transparency: Governance