Can OpenAI Beat Lawsuits With Its 'Hacking' Defense?

OpenAI is fending off half a dozen copyright lawsuits that allege the tech company illegally used copyrighted materials to train its flagship product, ChatGPT. The company's defense in these suits, which have the potential to shape the future of AI, is raising some eyebrows in the world of IP law.

OpenAI says the authors, news outlets, and other copyright holders who filed these suits engaged in "prompt hacking"—that, to get ChatGPT to spit out their books or articles, they had to manipulate the software by submitting thousands of prompts. And, to prove this, OpenAI's attorneys want the plaintiffs to turn over all the material they used to generate evidence for their suits, a request the plaintiffs say is excessively broad.

On this episode of our podcast, On The Merits, Bloomberg Law reporter Annelise Gilbert explains the strategy behind OpenAI's use of this defense and why the company may be trying to win not just in federal court but also in the court of public opinion. She also talks about what AI might look like if OpenAI loses these cases.

Do you have feedback on this episode of On The Merits? Give us a call and leave a voicemail at 703-341-3690.