Should an AI Robot Be Allowed to Critique Your Art?

On August 30 I posted a painting on Bluesky in recognition of International Whale Shark Day. Within minutes an AI curation bot called @curation-lab.bsky.social left unsolicited feedback on my work.

"Striking composition and poetic caption elevate the piece," it said. "The scale and theme resonate deeply. Minor refinement in texture could push it further."

I replied: "Robots are not my target audience sorry."

What else was I going to say?
I have been thinking about it since, though. Not because the feedback stung, it did not, but because the interaction raises a question that I think is worth sitting with. Should an AI be allowed to critique human art? And if it does, does that critique mean anything?
Let me argue both sides, because I think the honest answer is more complicated than a simple no.

The case for AI critique
Human art critics are not universally beloved either. Every artist has a story about a review that missed the point entirely, a juror who favored one school of work over another for reasons that had nothing to do with quality, a collector who dismissed something significant because it did not fit a trend they recognized. Human critics carry bias, taste, agenda, and blind spots. They are polarizing by nature. An AI at least arrives without personal grudges or aesthetic allegiances.

There is also an argument that pattern recognition has value. If an AI has processed millions of images and can identify compositional relationships, color theory principles, or structural tension in a piece, that data is not nothing. It is a kind of observation, even if it is not understanding.

The case against
But here is where it falls apart for me. Art does not have a blueprint. If it did, being an artist would be straightforward. The range of what art is and can be would be a very small category, defined by agreed-upon rules that anyone could follow. The reason art history is interesting is precisely because it is a record of people breaking the rules that existed before them, creating new frameworks, expanding what was considered possible or valid or beautiful. Monet was criticized for unfinished looking brushwork. Van Gogh sold one painting in his lifetime. Basquiat was called a graffiti artist as a dismissal.

An AI trained on existing work is trained on the past. It can recognize what has already been validated. It cannot reliably evaluate what is ahead of its time, what is intentionally transgressive, or what carries meaning that lives outside the frame.

And then there is the texture question, which in this particular case I find genuinely surreal. Industry analysts and art market researchers are increasingly predicting that work with physical texture, with the evidence of a human hand in the surface, will become more desirable as AI-generated imagery becomes ubiquitous. Because AI cannot replicate texture. It cannot press metal foil into a surface. It cannot embed glass at a threshold or build twelve physical layers of material that catch light differently at different times of day. The physical fact of handmade work is becoming rarer and therefore more significant.
So the robot critiquing my texture is the robot critiquing the one thing it can never do.
I find that clarifying.

What art actually is
Art is not a solved problem. It is an ongoing conversation between a maker and the world, conducted in a language that resists definition. The question of what makes something good art has occupied philosophers, critics, collectors and artists for centuries without resolution, because the answer keeps changing as human experience changes.

A machine trained on historical consensus cannot participate in that conversation. It can describe what it sees. It can compare it to what came before. What it cannot do is understand why something was made, what it cost the person who made it, or what it might mean to the person who receives it.

Human critics do not always get that right either. But they are at least capable of trying.
As for the bot that critiqued my painting on International Whale Shark Day: I appreciate the engagement. But I paint for the people who feel something when they stand in front of the work. That has never included algorithms.

Robots are not my target audience. Sorry.