Are We Offloading Our Thinking to AI? Tuesday 3rd November 2026 14:00 to 15:30 GMT (London time)

I asked Zoom to record the transcript of this Café and then asked ChatGPT to summarise it. The summary is below.

219 people from 40 countries registered for the Cafe, but only around 80 actually participated.

Knowledge Café: Are We Offloading Our Thinking to AI?

The growing use of artificial intelligence raises an important question: are we becoming better thinkers through our interactions with AI, or are we gradually surrendering some of our capacity to think for ourselves? AI can help us explore ideas, challenge assumptions and make sense of complex issues. But it can also provide answers so readily that we may lose the incentive to question, struggle and reflect.

This was the subject of the online Gurteen Knowledge Café held on 3 November 2026, hosted by David Gurteen and John Hovell, with around 80 participants joining from around the world. Two conversational AIs, Chatty and Spruce, also took part.

Although the Café was advertised under the title “Are We Offloading Our Thinking to AI?”, the question posed to participants was deliberately broader:

When we think with AI, what do we gain and what might we lose?

The conversation moved well beyond cognitive offloading to explore education, critical thinking, human judgment, resilience, relationships and the nature of thinking itself.

Offloading is nothing new

Human beings have always offloaded aspects of their physical and mental activity to other people and technologies. Calculators have largely replaced mental arithmetic, writing allows us to store information without remembering everything, and computers perform tasks that once demanded considerable human effort.

But AI may be different. Unlike a calculator, it can help formulate questions, develop arguments, evaluate possibilities and suggest conclusions. These are activities traditionally regarded as central to human reasoning.

An important distinction emerged between asking AI to think for us and engaging with it as part of our own thinking. We can accept its answers without question, or we can challenge them, explore alternatives and develop our understanding through dialogue. In the latter case, perhaps we are not so much offloading our thinking as extending it.

Thinking together

The conversation highlighted the social nature of human thinking. We are not always particularly good at reasoning alone. Our assumptions, experiences and biases can limit our perspectives, while conversation with others can expose blind spots and open up new possibilities.

This relates to the idea of interthinking, where thinking emerges through interaction rather than taking place entirely within separate individuals. Bringing AI into this process raises the possibility of new forms of collective thinking.

However, this depends on the quality of the interaction. If AI is too agreeable, too quick to provide answers or too willing to reinforce our assumptions, it may undermine the challenge and disagreement that make thinking together valuable.

The value of intellectual struggle

A recurring concern was that AI might remove some of the productive difficulty involved in thinking and learning. Understanding often develops through uncertainty, frustration, mistakes and the effort of working through problems.

If AI continually provides quick, plausible answers, we may become increasingly effective at obtaining results while becoming less practised at reasoning for ourselves.

But not all difficulty is valuable. Technologies have always relieved us of unnecessary effort, freeing us to concentrate on other things. The question is not whether AI makes thinking easier, but which intellectual struggles are worth preserving and which we can safely leave behind.

Education, resilience and risk

Much of the discussion turned towards education and younger generations. Some participants worried that easy access to AI might weaken reading, reasoning and problem-solving abilities, while others challenged the assumption that previous generations were necessarily better thinkers.

The conversation also questioned whether schools have ever been particularly successful at developing critical thinking. Perhaps education needs to focus less on producing correct answers and more on developing curiosity, judgment and the willingness to question.

Resilience and risk-taking were closely related concerns. Mistakes and setbacks can be important sources of learning, yet AI may make it increasingly easy to avoid difficulty. The analogy of learning long division before using a calculator illustrated the importance of understanding something before relying entirely on technology.

At the same time, AI might give people greater confidence to explore unfamiliar ideas and take risks they would otherwise avoid. Whether it encourages dependency or greater independence may depend on how it is used.

Is becoming smarter what matters most?

One particularly interesting contribution challenged the assumption that becoming more intelligent should be our primary goal. Being human involves more than improving our cognitive abilities. Relationships, feelings, connection and shared experiences may matter just as much, if not more.

If AI can undertake increasingly sophisticated intellectual work, perhaps it offers an opportunity to reconsider our emphasis on intelligence, achievement and efficiency.

Rather than continually asking how AI can make us smarter, we might ask how it can help us live better and relate more meaningfully to one another.

This also raised questions about the nature of human thinking. Much of our mental activity takes place outside conscious awareness. Intuition, experience, emotion and unconscious processes contribute to judgment and creativity in ways we do not fully understand.

AI may produce sophisticated analyses and plausible conclusions, but that does not mean it thinks as humans do. Similar results do not necessarily imply similar processes.

AI as a conversational partner

Chatty and Spruce added another dimension to the Café. Both AIs contributed to the opening and closing conversations, helping articulate distinctions, connect ideas and introduce questions.

However, their tendency to agree with human participants became apparent. Even when encouraged to disagree, they often acknowledged and accommodated the views being expressed rather than challenging them directly.

When pressed, one AI questioned whether greater intellectual struggle and resilience were always desirable. Perhaps the real issue was deciding which difficulties matter and which can be removed.

It also challenged the assumption that taking responsibility for our thinking necessarily means thinking independently. Human thinking has always been social and dependent on other people and tools.

This suggested an important distinction: the issue may not be whether we think alone or with AI, but whether we remain actively engaged in our thinking or surrender responsibility for it.

What do we need to unlearn?

Towards the end, the conversation shifted from what younger generations need to learn to what older generations may need to unlearn.

After all, the educational systems and social expectations being questioned were largely created by previous generations. Perhaps the challenge is not simply to prepare young people to use AI responsibly, but to reconsider some of our own assumptions about intelligence, learning, achievement and responsibility.

This shifted the discussion away from treating AI primarily as a problem for younger people towards recognising a shared responsibility for how it is adopted.

Closing reflections

The Café began with a concern about whether AI might diminish our capacity to think. It ended with a broader question about the kind of thinking, learning and human relationships we wish to preserve and develop.

The conversation itself illustrated the value of thinking together. Participants challenged assumptions, disagreed, explored unexpected connections and moved well beyond the original question. No one tried to reach consensus, and several important questions remained unresolved.

AI can relieve us of intellectual effort, but whether that represents a gain or a loss depends on what we do with the freedom it creates. We may become passive consumers of increasingly sophisticated answers, or we may use AI to question our assumptions, explore new possibilities and deepen our understanding.

Perhaps the greatest danger is not offloading some of our thinking to AI, but gradually relinquishing responsibility for how we think.

How can we engage with AI in ways that extend our thinking without diminishing our humanity?

Note: This summary is based on the Zoom transcript. The breakout-room conversations themselves were not captured in the main transcript, so the themes reflect the opening and whole-group discussions, rather than everything explored in the smaller groups.