Ask HN: The legal questions for AI Agent developers

What are the legal questions a product developer should think about that can/might be asked or products that are AI Agents for B2C use cases? Like the most fundamental is a READ operation: retrieving data from any platform that is a moat without intervention, followed by WRITE operations.

By Moats I mean API + Cost for example, most recently, Reddit's API pricing

While agents like ChatGPT have shown to have automated the querying of Freely available "shallow" knowledge. The next iteration of agents might be programmed by regular people to retrieve just about anything.

This is not a new/novel idea, but my question is if someone does manages to package this as a product and sells it at scale(say a phone maker), what might be the legal questions a product developer might face?