A new AI program is creating a buzz within the art world: Think of a phrase. Any phrase. Any collection of colorful, crazy concepts that you like. Research laboratory OpenAI’s new artificial intelligence program will create an art piece in less than a minute.

Introducing DALL-E 2: The DALL-E 2 program was created by sourcing an enormous amount of images and adding detailed captions to them, allowing the program to pull up and combine elements of images based on the inputs provided.

The concept is simple enough: You type in a phrase or a group of words into a search box, and the program creates a series of images from what you described. The full program still hasn’t been released to the public, but several people have been given access to experiment with it, and a simplified version, DALL-E mini, is accessible to all. Adobe Research’s Aaron Hertzmann has an article on his blog dedicated to all the artwork he has created from DALL-E 2, trying out different styles and techniques such as images in the styles of contemporary and modern artists, architectural images, classic computer graphics imagery, and portraits.

Enterprise gave it a try: We had some fun with DALL-E mini and typed “flower ducks” into the search box. Around 27 seconds we were met with nine pictures of ducks that seem to have taken on flower-like qualities in their coloring and texture. We weren’t super impressed with the outcome as the ducks’ appearance turned somewhat demented, but maybe our search phrase just wasn’t up to par. How about “coyote in a kilt”? Or “Donald Trump ballerina”? No, these are just terrifying.

Not everyone is a fan: CNet puts it best, calling the creations “nightmare fuel” — which we don’t think is overly harsh considering what we managed to generate. That being said, DALL-E mini is apparently a poor imitation of the all-powerful DALL-E 2, which seems to be far more sophisticated in the art it generates.

But DALL-E 2 is opening new conversations around how we define art: Can we consider algorithmically-generated images works of art? The fact that “creative AI” is receiving pushback from artists and audiences is “a sign of the power and potential of the new medium,” Stephen Marche writes for Lit Hub. “The most exciting promise of creative AI is that it runs in complete opposition to the overarching value that defines contemporary art: Identity. The practice itself removes identity from the equation,” he writes. Instead, creative AI is a recombination and reframing of other identities, with the possibility of choosing old artists or artistic styles and allowing technology to reimagine them based on any random phrase you come up with. “The power of creative AI is its strange mixture of human and other. The revelation of the medium will be the exploitation of that fact.”

It also raises questions about what it means to be creative when technology now has the ability to automate the process: The quality and realism of images created by DALL-E 2’s software is giving digital art a run for its money. For those who aren’t gifted with artistic ability but would like to see a certain image, these softwares can now pump out visually appealing results within seconds — a much faster process than hiring a human artist.

The controversy surrounding AI art is reminiscent of the outcry when photography emerged as a medium: When the first camera was created in the 1800s, it threatened everything the world believed about art. At a time where it was up to artists to capture the essence of a person in a portrait — a process that could take hours or days at times — the technology behind photography seemed to threaten their livelihood and the entire existence of art. Why paint a landscape when you can just take a photo of it? Until today, debates continue about whether photography can be considered art. Nonetheless, even with cameras being more accessible than ever, the continued importance of human-made art remains a testament to its necessity as a form of expression.

But in the same way as photography, creative AI should have a different approach to its development, argues Marche. AI should not be used to create art that people could make otherwise is the first point he makes. Meanwhile, creative AI should develop as more than just a showcase of the abilities of new technology. Creative AI could help the graphic design industry by offering tools that allow you to move things around in an image or conjure up an extra element — options that DALL-E 2 is working to provide, according to its website.

There’s also an ethical side to the equation — one which Google is currently working on: Since these AI ‘artists’ are designed by people, the risk that human biases creep into the algorithm are real. Google is trying to jump on the AI art trend with these concerns in mind, recently announcing Imagen — a text-to-image program “with an unprecedented degree of photorealism and a deep level of language understanding.” All the images the program produces have no people in them in case it reflects a Western bias and viewpoint. For this reason also, Imagen is not available for public use at the time, but this could change as the researchers develop more ethical code. “In future work, we will explore a framework for responsible externalization that balances the value of external auditing with the risks of unrestricted open-access,” Google researchers said in a paper (pdf).