That escalated quickly. George RR Martin and John Grisham are suing OpenAI. Why, and what could be the continuation? Let's not ask ChatGPT though. Let's think in an old-fashioned way. After all, our brain is just another embodiment of a self-learning neural network.
It’s natural why Mr. Martin is worried - why would we need a writer if we can ask chatGPT to write a book that perfectly suits our needs? Why would we need an artist, a composer, or an architect?
Let's consider a best-selling book example. Did the writer intend to make it popular, or was it an unstoppable inspiration? Many best-selling books are not so original. Let's face it: they are just compilations of the latest popular ideas, trends, and research results. And they are often "From the creators of ..." If you take "A Song of Ice and Fire", "Harry Potter", or any other bombing-popular long series, it's not surprising that fanfics can potentially overcome the original episodes. It's natural because the human brain is a neural network, that learns from the content consumed, just like ChatGPT. So, molest and monotonous content, fanfics, and sequel creation can be delegated to AI-powered language models.
What about real inspiration? The problem is that the mechanisms of creativity and resourcefulness are not yet well understood. It’s obvious, that humans learn from the content consumed, but different people learn differently from the same content. Although there are some guidelines for creativity improvement, it’s to a high extent bare marketing and prevarication.
Let’s say someone creates a fanfic based on their demand. Would that fanfic fulfill the demand of the whole fandom as well? If yes, this could mean that the existing content is exhaustive, and the writer’s input might not be needed anymore. Pretty tough, but that’s only a fact. Creativity and art are one thing, but earning funds and intellectual property is way different, so let’s keep it applicable.
Do you remember the Turing Test criterion?
If a machine can engage in a conversation with a human without being detected as a machine, it has demonstrated human intelligence.
Let me rephrase Alan Turing’s thought in the opposite direction:
If a human cannot be distinguished from a machine, is it any better than a machine?
I believe that before the mechanisms of human creativity are well-studied, advanced language models are no threat to creativity as such. It’s a bit different for some forms of business that freeload creativity though. Even if those mechanisms were studied, the outcome of the generative model depends on how creative the request is, so it would still depend on art and fresh thinking.
One thing to learn from the current situation is that if you are producing content that looks like ChatGPT’s text, if your speech is too generic, maybe it’s better to be kept untold to save energy and everybody’s time because we already have AI for that. Keep it genuine, technology is only a tool. Scientific progress is unstoppable, the question is only can you use it for good, or not.