We want to preface this one and say that we think Generative AI is really cool and already have integrated some of these tools into our daily lives (See e.g., Robot in Suit above). That said, there are quite a few legal issues we are keeping our eyes out for (privacy, intellectual property, hallucinated fake case law, etc.).
Sitting in the middle of the bingo card of potential legal issues is training on copyrighted work. An AI writes a book but only after training on a bunch of other books. Is that copyright infringement? Well, that’s already being tested in two recently filed lawsuits: one by Comedian Sarah Silverman and the other by a group of authors.
Now let’s get a little bit broader. What about social posts, chats, comments, replies, searches, keystrokes, mouse clicks/movements, signals, or browser activity? That’s the subject of a new class action filed against OpenAI and Microsoft.
In a nutshell, the named plaintiffs allege that OpenAI/Microsoft have unlawfully scrapped potential class members’ data and used it to train their AI models.
Interestingly, we are seeing these allegations for a second time; a very similar class action was also filed against OpenAI and Microsoft just a few months ago. Will be interesting to see if more of these class actions pop up.