The Curious Case of Dr. Seuss & Generative AI: Navigating the Hype Cycle
Academic and public fascination with generative models stems from their remarkable multi-modal capabilities, encompassing text, images, and music that emulate human creativity. However, it's crucial to balance our excitement with a dose of reality. The intriguing example of how major genAIs represent Dr. Seuss serves as a poignant reminder of this necessity.
Today, I had planned to commemorate the birthday of Ted Geisel (image above), renowned as Dr. Seuss, with a Substack poll to determine which Generative AI crafted the finest Dr. Seuss-style short story about his birthday, interwoven with reflections on the emergence of machine learning, deep learning, and generative AIs.
Utilizing ChatGPT 4.0, Mistral, Anthropic Claude, and Google Gemini, I sought to channel the whimsy of Dr. Seuss into a narrative exploring the complexities and potentials of AI technologies. However, the outcome was disappointing.
The prompt I used was as follows:
Using the Dr. Seuss-style of writing create a 10-sentence short story celebrating the birthday of Ted Geisel (aka Dr. Seuss) and the power and risks of AI, machine learning, deep learning and generative AIs
The outcomes were disappointing, with several notable observations:
The genAIs didn't strictly adhere to the instructions on length and failed to seek clarifications.
ChatGPT's response deviated from the others, lacking the whimsical charm associated with Dr. Seuss.
Gemini initially highlighted copyright concerns regarding Dr. Seuss but still provided a short story.
Despite efforts, even the stronger responses from Gemini and Llama-Groq felt constrained within the text.
Overall, the content fell short, lacking the imaginative spark and depth synonymous with Dr. Seuss's work.
Instead of compiling a lengthy 5-page blog post detailing all the results, I have opted (with advice from a reader) not to include them. Instead, I have included the prompt for readers of this blog who may wish to replicate the outcome themselves – even if results will vary.
The disparities bullet-pointed above don't signify AI failure but rather illuminate the current state of technology. They underscore a crucial reality: the excitement surrounding AI surpasses its actual developmental stage. These challenges highlight the present limitations of AI models, which, despite progress, struggle to fully comprehend and replicate human creativity and emotion.
This scenario echoes the Gartner Hype Cycle, depicting the trajectory of technology adoption. It suggests that new technologies often experience a period of inflated expectations before confronting disillusionment with challenges. GenAIs journey, particularly in creative endeavors like mimicking Dr. Seuss, follows this pattern.
The intersection of Dr. Seuss and genAIs isn't a story of failure but a paragraph in the chapter in the ongoing narrative of AI-technology advancement. It serves as a rallying cry for researchers, developers, and enthusiasts to push AI's boundaries while acknowledging its current constraints.