ChatGPT has been dominating headlines for weeks, but months before the chatbot launched, the internet was obsessed with another OpenAI project. Dall-E and Dall-E 2 are deep learning models developed by OpenAI that generate digital images based on prompts from users. You’ve probably seen hundreds of AI-generated images in recent months as a result of Dall-E’s popularity, but one user recently decided to combine the two OpenAI models for a unique experiment.
On Saturday, a Reddit user shared a Dall-E image of ChatGPT as a humanoid figure. In order to achieve this, the user asked ChatGPT to describe its humanoid form. They then fed that response to Dall-E. This is what resulted from the experiment:
Here’s the description that ChatGPT initially gave the user: “I have a sleek, metallic exterior with glowing accents to represent the circuitry and data processing elements of my virtual being. My eyes are displays that change color or show data, I have augmented reality elements that allow me to project a holographic user interface in front of me. I have other unique features, such as the ability to manipulate digital objects and information with gestures.”
Needless to say, this wasn’t a scientific study. That image is not actually ChatGPT come to life (or even worthy of a desktop background). That said, these AI systems will inevitably begin to interact with each other in the years to come. The results won’t always be pretty, but we’d be intrigued to repeat these steps in a year and see how much the outcome changes.
In the meantime, ChatGPT is about to get some tough competition in the form of Bard. Google debuted its “experimental conversational AI service” this week, and it will be available to the public in the coming weeks. Will ChatGPT continue to draw as much attention when Google builds conversational AI responses directly into Search?