AI: From the Engine Room

Hallucination is a Feature, Not a Bug

Vendors frame hallucination as a bug they're fixing. It's actually inherent to how these systems work, which changes the governance question entirely.

Part 10 of 13 in AI: From the Engine Room

How Text Generation Works

Language models are next-token predictors. Given a sequence of text, they generate the most likely continuation based on patterns learned from training data.

Here’s the important part: ‘most likely’ doesn’t mean ‘true.’ The model doesn’t have a concept of truth. It has a concept of plausibility.

The model doesn’t have a concept of truth. It has a concept of plausibility. Hallucination isn’t a malfunction; it’s inherent to probabilistic text generation.

Why It Can’t Be Eliminated

Hallucination is inherent to probabilistic text generation. Mitigation strategies can reduce frequency but not eliminate it.

The Governance Implication

If hallucination can’t be eliminated technically, it becomes a governance problem: how do you manage systems that will sometimes confidently produce incorrect outputs?

Instead of ‘how do we fix hallucination?’ the better question might be ‘for which use cases is this error rate acceptable?’

The Takeaway: Hallucination is a property of probabilistic text generation, not a bug being fixed. Governance is about managing that property, not waiting for it to disappear.