OpenAI has licensing agreements with some content providers and allows webmasters to block its web crawler from scraping their site for training data. It also allows artists to "opt out" of and remove their work from the data sets used to train its image-generating models, though some artists have criticized this process as burdensome. Deng stated that artists should have more agency in the creation and use of generative AI tools, but did not specify how this might be achieved.
Key takeaways:
- Peter Deng, VP of consumer product at OpenAI, avoided answering whether artists whose work was used to train AI should be compensated, during a discussion at SXSW.
- OpenAI and other generative AI vendors argue that their use of public data for training AI systems, without compensating or crediting artists, is protected by the legal doctrine of fair use.
- A class action lawsuit has been filed by artists, including Grzegorz Rutkowski, against OpenAI and its rivals, arguing that these AI tools replicate artists’ styles without explicit permission and without payment.
- Deng believes that artists should have more agency in the creation and use of generative AI tools, but is unsure of what form that might take.