Accuracy Gap
The difference between how accurately AI and experienced people make the same call. The site's task and decision article reports AI at around 65% in ambiguous, judgement-heavy situations against roughly 85% for experienced humans, narrowing to 75 to 80% against 85 to 90% on structured predictions such as candidate shortlisting.
What it means in practice
The task and decision article treats the gap as the reason for its approach, not an afterthought. A system 20 points behind a person on the hardest calls and 10 points behind on the easiest is, as the article puts it, two different tools doing two different jobs. At 75 to 80%, AI output is useful as a second opinion that ranks or flags cases for a person to look at first. It is not close enough to stop checking.
How we use it
The article’s delegation ladder is where the gap is applied. Anything touching hiring, dismissal or a legal or safety obligation stays off the ladder however capable the tool looks, and the article says those categories map closely onto what counts as a high-risk AI system. Keeping a human in the loop is how the remaining Assist and Recommend steps stay checked.
FAQ
Questions about Accuracy Gap
How Accurate is AI compared with an experienced person?
It depends on how ambiguous the call is. The task and decision article cites a meta-analysis putting AI at around 65% in ambiguous, judgement-heavy situations against roughly 85% for experienced humans. On structured, data-rich predictions such as candidate shortlisting, it reports 75 to 80% for AI against 85 to 90% for people.
What should a Business do about the Accuracy Gap?
Sort each step of a process by how much to delegate, rather than approving the chain as a whole. The article's ladder keeps ambiguous or judgement-heavy steps at Assist or Recommend with a named person accountable, and lets high-volume, reversible, rule-based steps move to Execute.
What it means for a Business
The gap is a reason to sort each step by how much authority to hand over, not a reason to distrust automation. On judgement-heavy calls AI can support a person's decision without being trusted to make it alone.
Ready to Put This to Work?
Tell us where your team is with AI and we will tell you honestly what would make the biggest difference.