Artificial Stupidity (Eno)

In the AI Story

Eno's observation about artificial stupidity is grounded in a broader claim he has made across decades: that the most interesting uses of any technology are the uses the technology was not designed for. One of the things artists are interested in for technology, he has said, is the things that they do that they're not supposed to do. The dominant texture of any era is really captured in the shortcomings of those technologies. The distorted electric guitar emerged from the shortcoming of amplifiers pushed beyond their design parameters. The sound of early digital recording emerged from the shortcoming of low bit depths. The characteristic grain of early photography emerged from the shortcoming of early emulsions.

Applied to AI, the principle reframes fabrications, hallucinations, and confident errors from failures to be corrected into material to be examined. When Claude produces an elegant but factually incorrect reference, the conventional response is to treat this as a bug. Eno's response is to treat it as a feature — not in the sense that the error is desirable, but in the sense that the error reveals something about the system's processing that the correct output would have concealed. The mistake carries information the accuracy cannot.

The concept connects to the Oblique Strategies principle of honoring the error as a hidden intention. But artificial stupidity operates at a different scale. The Strategy addresses individual errors; artificial stupidity addresses the systematic peculiarities of a specific class of tool. Every medium has its characteristic stupidities — tape compression, analog saturation, digital quantization noise — and skilled practitioners across history have learned to work with these stupidities as palette rather than against them as problems. AI's stupidities are different in kind but identical in structural role: they are what the tool does that it was not designed to do, and they are where the creative opportunity lives.

The practical implication is that AI should not be used exclusively in its mode of maximum competence. The practitioner who seeks only reliable, correct, specification-matching output misses the dimension of the tool that has generative potential. The practitioner who learns to solicit the tool's weird mistakes — who asks questions the tool will answer strangely, pushes it into territory where its processing becomes visibly idiosyncratic, seeks the seams where the machine's alien cognition differs most dramatically from human thought — has access to material no other tool can provide.

Origin

The phrase artificial stupidity surfaced in Eno's interviews about AI in 2024 and 2025, particularly in conversations with the BBC and The Guardian. It builds on earlier formulations — capitalize on the shortcomings, the dominant texture of any era is captured in the shortcomings of technology — that Eno has used for decades to describe his approach to technological tools generally.

Key Ideas

Shortcomings are material. Every medium has characteristic stupidities that skilled practitioners learn to exploit rather than correct; AI is no different.

The error is idiosyncratic. AI mistakes are not random noise; they reflect the specific architecture of the system, and their patterns are what makes them useful.

Competence is not the interesting mode. The tool's correct output is the least distinctive use; the stupid output is where the tool's character becomes visible.

Seeking the error requires discipline. Most interactions with AI suppress its stupidity; eliciting the productive mistake requires prompts and contexts the default use does not encourage.