Article

A Human Future Needs Room to Challenge the Machine

1 November 2026

Imagine being told that a decision affecting your life is correct because a system has followed its rules. You explain that something important has been missed. The answer comes back more clearly, perhaps more politely, but nothing changes. Somewhere in this arrangement, your ability to speak has become detached from anyone’s obligation to listen.

This is one useful test of the promise to keep humanity central to artificial intelligence: can the person affected challenge the purpose of a decision, as well as point out an error in its calculation?

A Council of Europe media advisory for Web Summit 2026 announces a discussion involving a European human rights chief and a Vatican AI ethicist, framed around humanity’s place in AI development. The announcement establishes a subject for discussion, rather than an outcome. Its subject deserves attention beyond the conference, because agreement that technology should serve people leaves a considerable difficulty untouched. People disagree about what service means.

An organisation might want to process applications faster. An applicant might need someone to understand why their circumstances do not fit the form. Both concerns can be reasonable. Yet if speed becomes the governing measure, the awkward case begins to look like a defect in the person presenting it.

Before asking whether a machine performs a task accurately, we need room to ask whether the task has been defined adequately. That question cannot belong exclusively to whoever buys or builds the system.

There is an obvious complication. Every decision cannot become an unlimited negotiation. Public services have finite resources; other people are waiting. Human discretion can also be careless, prejudiced or inconsistent. There is no reason to romanticise the desk simply because someone sits behind it.

Still, a workable limit on appeals is different from designing away the possibility of reconsideration. A meaningful challenge needs to reach someone with the authority to change the result, including when the system has operated exactly as intended.

For an evolving philosophy concerned with how we live together, this reaches beyond technology. We inherit rules, grow accustomed to them, then forget that they contain choices. AI makes it especially important to keep those choices visible. Otherwise, we may spend years correcting the machine’s answers while leaving the question it was asked beyond anyone’s reach.