A question, not an answer
BBC Business has put a question at the centre of modern lending: should artificial intelligence be the judge of a person's creditworthiness? The article, published on 20 November 2026, frames the issue around three everyday products — a mortgage, a car loan and a mobile phone contract S1.
Creditworthiness is a lender's judgment about whether a borrower will repay. Errors cut both ways. Approve someone who cannot pay and the lender absorbs a loss; reject someone who could have paid and a household is shut out of a home, a car or a phone contract. Because those decisions shape access to essentials, who makes them — and how — is a money question, not only a technology one.
Why the framing matters
The BBC presents the answer as open rather than settled S1. That is the notable part. There is a difference between a system that scores a file and a system whose score is the last word. A flag passed to a human for review and a verdict delivered without one are not the same product, even if both are described as AI.
The three examples also sit at very different stakes. A mortgage is typically the largest borrowing a household takes on, a car loan is smaller and shorter, and a phone contract is small and routine. Any answer to the BBC's question would have to hold across all three S1, which is a demanding test: a method that works for a low-value, fast-turnaround decision may not be suitable for a 25-year secured loan.
What the source does not settle
The article as supplied raises the question without giving figures, named lenders, studies or a regulatory position S1. Readers should therefore treat it as an invitation to debate rather than evidence that AI is, or is not, already deciding loans. Claims about accuracy, bias, cost savings or turnaround times in this area need their own sourcing, and none is present here.
What the piece does establish is the framing: the subject is machine judgment of personal credit, applied to mainstream consumer products S1. Everything beyond that is still to be argued.
The trade-offs worth naming
Three tensions sit under the headline question. The first is speed against explanation: a fast decision is of little use to a borrower who cannot understand why they were refused. The second is consistency against circumstance: rules applied identically to everyone are easy to defend and sometimes wrong for an individual. The third is scale against accountability: the more decisions a system makes, the harder it becomes to trace a single one back to a person who can answer for it.
What to watch
- Whether lenders can explain a refusal in terms a borrower can challenge.
- Whether a human stays in the loop for mortgages and other large loans.
- How regulators treat automated credit decisions, and what evidence they ask for.
- Whether complaint and appeal routes keep pace with the speed of automated systems.
The question is not confined to one market. Any lender that scores borrowers, in Nairobi or London, faces the same tension between speed and fairness — and the same difficulty explaining a "no" to the person who receives it.
- Should AI be judging our credit worthiness? BBC Business · bbc.com