Generative AI models like ChatGPT, GPT-4, and DALL-E 3, which are trained on millions to billions of data from public websites, can potentially land users in trouble if they regurgitate a mirror copy of a training example. A recent survey found that intellectual property was a major concern for companies using generative AI. OpenAI now joins IBM, Microsoft, Amazon, Getty Images, Shutterstock, and Adobe in pledging to indemnify generative AI customers over IP rights claims.
Key takeaways:
- OpenAI has announced a new program, Copyright Shield, in which it will defend businesses using its products against copyright infringement claims and cover their legal costs.
- The protection applies specifically to customers using the "generally available" features of OpenAI’s developer platform and ChatGPT Enterprise, but not to all OpenAI products.
- Generative AI models, which are trained on a vast range of data from public websites, can potentially cause copyright issues if they produce a mirror copy of a training example.
- OpenAI joins other companies like IBM, Microsoft, Amazon, Getty Images, Shutterstock, and Adobe who have also pledged to indemnify generative AI customers over IP rights claims.