AI engines are getting smarter and smarter. They can do more and more of our thinking for us – so we are increasingly delegating complex thinking tasks to them. One of the results of this is that the more we use AI, the more we may unconsciously be dumbing ourselves down. Like any other muscle, our brain needs exercise to maintain its fitness.
A recent article by Sarah O’Connor in the Financial Times (21 April 2026, sarah.oconnor@ft.com) draws parallels between two other addictive behaviours – smoking and social media. As the dangers of smoking have become increasingly apparent, a class divide has emerged in who does and doesn’t give up. Those at the bottom of society are most likely to remain addicted. A similar phenomenon related to alcohol in the late 19th century. There is also evidence that the those at the bottom are less likely to engage in healthy exercise – health is a class issue.
The push-back against social media addiction that is gaining traction in developed economies, suggests O’Connor, may already be creating a similar class divide. Those least equipped to resist constant immersion in social media will come predominantly from the least advantaged groups in society.
So, will the same be true of AI engines? Will the least advantaged become the most dependent on AI to organise their lives? When does organising become controlling?
These are not just philosophical questions. The quality of life for the next generations will depend on how well we address these challenges.
©️David Clutterbuck, 2026