DeepSeek and the boundaries of intellectual property rights

We are going through a very interesting phase where the boundaries of intellectual property rights seem quite blurred.

OpenAI claims that DeepSeek used ChatGPT to train its AI model, DeepSeek R1, which amounts to IP theft. Interestingly, OpenAI itself used copyrighted materials from all over the world to train ChatGPT, disregarding concerns raised by many regarding intellectual property rights violations. In fact, there is a high possibility that Suchir Balaji was murdered because he went public with his IP rights violation allegations against OpenAI.

In short, DeepSeek allegedly stole something (from OpenAI) that did not belong to OpenAI in the first place. Coming back to what matters to us as academics, we are handing over intellectual property rights to Elsevier et al on a platter. Can AI come to the rescue?