AI at work raises a harder question than accuracy: who owns the decision?

When AI influences decisions about someone’s career, can a manager defend the outcome if they cannot explain how it was reached? Employment lawyer Mandy Laurie argues that accountability, not just accuracy, is a central challenge for organisations using AI at work.

Watch Desk analysis

What happened

In an opinion article for The Scotsman, Laurie considers AI-influenced decisions such as promotion, redundancy and performance assessments. Her concern is that a leader may trust a system’s recommendation yet struggle to interrogate its reasoning or explain it to the person affected.

Laurie argues that responsibility can become blurred between human decision-makers, software providers and governance processes. She also raises the possibility of “moral injury” when leaders are expected to endorse outcomes they cannot satisfy themselves are fair. These are arguments about the risks of workplace AI, not evidence that a particular employer or system has made an unlawful or unfair decision.

Why it matters

A decision can be statistically sound and still leave an employee asking a reasonable question: why me? If the answer is hidden behind a system no one at the organisation can meaningfully challenge, the promise of human accountability starts to look rather theoretical.

That makes explainability a practical management issue, not merely a technical nicety. Organisations need to know who can question an AI-influenced recommendation and who remains answerable for the final decision. Efficiency is a thin comfort if nobody can explain an outcome when it matters.

Our read

Laurie’s strongest point is that generating a recommendation and earning the right to act on it are not the same thing. Accuracy and legal compliance matter, but they do not settle whether a decision-maker can justify the process to the person whose livelihood is at stake.

This is an opinion piece, not a report of a new rule or a documented workplace case. Still, it puts a useful test to employers: if a manager cannot explain or challenge an AI-influenced decision, has the organisation kept meaningful human responsibility, or just kept a human in the meeting?

What to watch

• Whether employers set clear accountability for decisions influenced by AI.
• Whether workplace AI systems let decision-makers examine and challenge the reasoning behind recommendations.
• Whether staff can contest those decisions in practice, not only on paper.

Discussion spark: If a manager cannot explain an AI-influenced decision about someone’s job, should they be allowed to act on it at all?

Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.