OpenAI also mentioned that it is continually making its systems more resistant to adversarial attacks to regurgitate training data. The company collaborates with news organizations to explore new opportunities, and has pursued partnerships to deploy its AI products to support reporters and editors, train its AI models on additional content, and display news publishers’ real-time content with attribution in ChatGPT. An example of such a collaboration is a licensing deal with media company Axel Springer announced on Dec. 13.
Key takeaways:
- The New York Times has sued OpenAI and Microsoft, alleging copyright infringement and claiming the tech companies used its content without permission to develop their AI products.
- OpenAI has responded saying no single source, including The New York Times, makes a meaningful contribution to the training of its models and any misuse of publishers’ content is not typical or allowed user activity.
- OpenAI is working on making its systems more resistant to adversarial attacks to regurgitate training data and has made progress with its recent models.
- OpenAI has been collaborating with news organizations, including a licensing deal with media company Axel Springer, to deploy its AI products to support reporters and editors, train its AI models on additional content, and display news publishers’ real-time content with attribution in ChatGPT.