What If SkyNet Engages?

The recent interest in and explosion of AI chatbots poses an interesting question – can an AI write a novel?  And what happens if it can?

The first question is whether or not it can write a novel.  By that, I don’t mean that it puts together a few phrases it learned from a writer’s preferences, but can it generate something unique and thought-provoking?  I believe that’s doubtful in the current climate since these AIs aren’t truly intelligent on their own.  To begin with, they’re mostly algorithms designed to elicit a predictable response based on tendencies.  Second, and most important in my opinion, is that all the algorithms in question are programmed by people, and people have biases and tendencies.  This isn’t some child who will initially mimic his or her parents but come o form independent thoughts, but simply a computer program that will perform according to the biases of the person who wrote it(the true danger of AI is when someone you disagree with programs an AI to enforce rules you think are horrible).

The second question is much more intriguing.  We can probably prod an AI to come up with a story, but who owns that story?  What are the copyright implications of such a thing?  Current law doesn’t recognize AI as having any rights of its own, so would the writer of the algorithm own the story?  Or would it simply be part of the public sphere immediately, and thus subject to people borrowing from it(or outright stealing everything the AI “wrote”)?  I don’t believe machines or programs can own copyrights.

We’ve had innumerable sci-fi stories about AI rising up to take control, but we have very few that describe what AI would do in society beyond “kill all the humans.”  Now that they can (kinda) compose things and hold pseudo-conversations, maybe it’s time to begin thinking about where this is going.