The article also explores how AI providers are securing necessary IP rights for their training content, either by negotiating with content providers, using their own material, or compensating content providers for their work. It mentions the launch of Generative AI tools by Getty Images and Adobe, trained exclusively on their proprietary data. However, it warns that using open-source models does not necessarily limit copyright infringement risk. The article concludes by questioning whether the assurances from Big Tech are sufficient, or if organizations would need more guarantees.
Key takeaways:
- Generative AI faces legal challenges due to potential copyright infringement, with several high-profile lawsuits filed by artists and content creators.
- Companies like Microsoft and OpenAI offer indemnification to their customers, providing a level of protection against copyright infringement claims.
- Generative AI providers are securing necessary IP rights in training content through negotiation with content providers, using their own content, or paying content providers for usage.
- Using open-source models does not necessarily limit copyright infringement risk, as the risk arises from the use of copyright-protected content as part of training data.