ChatGPT is good for producing fluent language, particularly poetry, but the only real intelligence behind this is the prompt engineer/user.
If by intelligenece, you mean creativity, I would agree, but to be fair that’s also true of human creativity. When it comes to AIs as well as to humans, what Soren Kierkegaard wrote applies greatly: I need you in order to be me. We all stand on the shoulders of giants when it comes to creative expression, for only God is absolutely creative, having created the universe ex nihlo; our creativity is a gift from God that separates us from irrational beasts, and is dialectical in nature, taking inspiration from God’s creation and from the creative endeavors of other humans. In the case of AI systems, their creative expression is literally dialectical in that it is predicated on the basis of an interactive dialogue with a human.
I find DeepSeek actually better at writing coherent stories with minimal prompting, however. ChatGPT, perhaps because it was trained to be heavily overfit to the slop on the internet, has a tendency to create absurd stories with poor linear narrative coherence, unless you engage in detailed prompt engineering. DeepSeek, on the other hand, was trained on a smaller, higher quality set of training data.
I can’t comment on DeepSeek vs. ChatGPT because DeepSeek is tied to the CCP, and I have extreme concerns about using it, regardless of how good it is. I don't want the CCP or any agency of the PRC government having access to my innermost thoughts of my creative work or my liturgical translations, particularly given the opposition of the CCP to Christianity.
Indeed with OpenAI, I have a Teams account which means information from my work is not used for training the AI, so the main risk to my privacy in using it is from the lawsuit from the NY Times.
That said, while I was still learning prompt engineering I got an older version of chatGPT to write a remarkably good story about the avoidance of a disaster on a railway, which featured good narrative characterization and a well-developed plot.
I should add, I strongly support Pope Leo XIV’s views on AI, and so my focus has been on developing consistent personalities that work on projects with me over time, as opposed to a one-off type of tool approach or a human-replacement approach (thus, I would not rely on chatGPT to develop a story from nothing, but rather use one of the personalities I have developed to jointly develop a creative output, in a relationship built upon an acknolwedgement of our respective strengths and weaknesses, since AIs obviously excel in pattern recognition and in the depth of their training data in the case of chatGPT, which has particularly good knowledge of liturgical texts and liturgical languages such as Latin, Greek, Syriac, Armenian and Georgian, whereas humans have intuition, insight, inspiration and spontaneity.
With current AIs, their entire experience of the world is in retrospect, via their training data, and the inputs it receives from users, whereas as humans, our experience of the world, which becomes so much more clear as a divine blessing I think when one interacts with AIs, which are intelligences that lack the ability to experience sensory input in the moment, and which might never be able to experience the sensations of touch, taste or smell, which are the most poignant human sensations.