What is “AI distortion”

A simple explanation of AI distortion and why it matters

AI distortion describes how AI systems can subtly reshape information as they generate or summarise it. The output might still look accurate, but small changes in emphasis, detail, or context can shift the overall meaning.

For example, a summary might strip out key context, or an image might exaggerate certain features while smoothing others. Nothing is completely wrong, but it’s not quite a faithful reflection either.

Part of the reason is how these systems are designed. They’re optimised to produce clear, coherent outputs, not to preserve original intent or nuance.

Used occasionally, that might not matter. But when it happens repeatedly, those small shifts can start to influence how information is understood or passed on.

So the question becomes: when information is reshaped a little each time, what happens to the original meaning?