OpenAI has quietly released text-davinci-003, a brand new entry in the artificial intelligence GPT-3 language model family, which claims it can handle more complex prompts to produce longer output.
However, as reported by Ars Technica (opens in new tab)enterprising users Playground (opens in new tab)the free offering of GPT-3, quickly found that the new model is even more adept at producing poetry and lyrics.
Hacker news commentators found it (opens in new tab) that it could write poems about Einstein’s theory of relativity, then rewrite them in the style of the romantic poet John Keats. While Professor Ethan Mollick, a professor at the University of Pennsylvania’s Wharton Business School, waxed lyrical about the possibilities in a series of tweets (opens in new tab).
GPT-3 art generation and the future
The improvements in GPT-3’s understanding of rhyme and meter are likely the result of filling it with even more reference material. The Github repository for GPT-3 (opens in new tab) recognizes that it derives its massive corpus from thousands of datasets.
Have lovers noted (opens in new tab) that previous iterations of GPT-3 have had some cursory knowledge of rhyme schemes, but this latest step forward is a sign that this AI writer now has a complex understanding of meter and could eventually compose artistic works of his own.
While all these developments are exciting, they do raise the question of how human artists, writers and – er – journalists can coexist with technology that is becoming less and less “bleeding” by the day.
Some fear that AI, which can write and rewrite faster than any human, will strip them of paying work. AI-powered generation and manipulation of text is nothing new, with tools like Language is a virus (opens in new tab) and the GPT-3 powered InferKit (opens in new tab) have offered some version of these features for some time.
While it’s true that AI text generators (and art generators like DALL·E) take a lot of work out of creativity, humans still have to create the prompts. And when it comes to the ability to regenerate certain parts of an output as dictated by a person (“paint in (opens in new tab)”, which both DALL E and the alternative Stable Diffusion are both capable of), this too is a human-driven process.
So instead of seeing the latest crop of truly competent AI generation systems as a threat to human creativity, we might think about how it can accommodate and interact with them.
This can be as a source of inspiration, making artistic processes accessible to more people or creating new human/AI hybrid processes.
And if you still feel threatened, consider this: if the content you produce is replicable by an AI system… wouldn’t you rather be making something else? A new and improved GPT-3 (or GPT-4, which is rumored to be coming) could give you more time to do that.