A Letter to a Generation Growing Up Beside a Machine

Alan Turing opened his famous paper by saying he would consider the question 'Can machines think?', then set it aside for a test that could actually be run: an imitation game in which an interrogator tries to tell the machine from the human. He predicted that, in about fifty years, computers would play it so well that an average interrogator would have no more than a 70% chance of identifying the machine after five minutes. That was an estimate by its author, not an established fact, and the unseen is in God's hands.

Five years later, in 1955, John McCarthy, Marvin Minsky, Nathaniel Rochester and Claude Shannon wrote a proposal for a summer study at Dartmouth in 1956 involving ten researchers. It proceeded from the hypothesis that every aspect of learning, or any other feature of intelligence, can be described so precisely that a machine can be made to simulate it. Ever since, we have been asking what the machine can do. The question that concerns those growing up in its company is different: what do we keep in our own hands?

The danger is not rebellion

The first worry is that we hand the tool what builds our own faculties. In a survey of 319 knowledge workers (Lee et al., 2025), higher confidence in the tool was associated with less critical thinking, while confidence in one's own ability on the task was associated with more. The data are self-reported, and correlation does not establish cause. But the practical lesson stands on two feet: do not boycott the tool, and do not let it think for you all the time. We have addressed the educational side in the article 'The End of the Age of Rote Learning?'.

And the algorithm that shapes attention?

It is often said that feed algorithms shape our desires and feed polarisation. The evidence is less conclusive. In a field experiment during the 2020 US election, consenting users were moved from the default algorithmic feed to a reverse-chronological feed for three months. They spent less time on the two platforms and saw more political content and more untrustworthy sources, but polarisation on issues and in feelings did not change significantly over that period.

That does not mean feeds are innocent; three months is short, and one experiment does not close the question. But it invites precision: we guard our attention because it is precious, not because we have shown that every algorithm corrupts.

A measure of success

A technology is measured by what it does to us, not by what it achieves. A tool that lets a doctor be more present with a patient, and a teacher listen more to a pupil, is better than one that replaces them merely because it is cheaper. If productivity benefits some while the weakest, those furthest from the table of technology, remain outside its benefit, the success is incomplete. This is a value-based measure proposed by the editors, not a measured result; whoever adopts it asks of every technology: whom does it benefit, and who is left outside it?

Five counsels

• Ask of every application: does it widen human dignity and capability, or narrow them?
• No momentous decision in medicine, law or education is handed over unless a human keeps the right to review and to hold to account (see the article 'Principles of Responsible AI').
• Question what you are given before you swallow it, and remember what we said about bias in the article 'Algorithmic Bias'.
• Think through the problem before you ask the tool, and leave your mind an effort that no one can do for it.
• Keep 'why?' for yourself: the tool helps with 'how', while purpose remains with whoever bears responsibility.

The mind is a trust, and thanking God for it means using it when the tool lets us put it down.