Judgement in the Age of AI … You Can’t Always Get What You Want
Why Philosophy Matters More Than Ever (Part Two of Two)
In the previous post we reflected on the slightly curious sight of philosophers appearing inside AI firms. At first glance the development seems faintly incongruous. Why would organizations built on mathematics, engineering and enormous quantities of computing power need abstract thinkers.
Yet the more one thinks about it, the less surprising it becomes. Artificial intelligence is forcing society to revisit questions that philosophy has been asking for centuries. Questions about intelligence itself. Questions about responsibility and moral agency. Questions about how humans should live alongside increasingly powerful technologies that are capable of influencing decisions, shaping knowledge and, in subtle ways, participating in the thinking process. As Socrates observed, “wonder is the feeling of a philosopher, and philosophy begins in wonder.” Artificial intelligence has a way of producing exactly that reaction.
That last point is perhaps the most significant. Artificial intelligence is not simply another wave of technological innovation in the mold of the smartphone or the spreadsheet. It is, increasingly, a technology that interacts with cognition itself. It participates in the production of ideas, the interpretation of information and the formation of arguments. And once technology begins to influence the process of thinking, the concerns of philosophy inevitably return to the foreground.
For most of human history, thinking was an activity that happened primarily inside individual minds. A biological process. People read books, wrote arguments, analyzed information and debated ideas with other people. Even when collaboration took place, cognition remained a distinctly human activity. Tools might assist the process, but they did not actively participate in it. Artificial intelligence changes that arrangement.
Increasingly, knowledge work involves a hybrid process in which humans and machines think together. AI systems generate summaries, produce drafts, surface patterns in data and suggest connections between ideas that might otherwise remain hidden. Humans then interpret those outputs, evaluate them, reshape them and ultimately decide which of them are worth pursuing.
What emerges from this interaction is something that begins to resemble a new form of collective intelligence. The term has traditionally been used to describe the way groups of people combine their knowledge and perspectives to produce outcomes that exceed the capabilities of any single individual. Organizational theorists and management scholars have explored this phenomenon for decades, noting how networks of individuals, institutions and digital tools can combine to create systems of distributed intelligence.
Artificial intelligence extends that idea further still. The “collective” in collective intelligence may increasingly include machines as well as humans. That development has obvious benefits. Machines are exceptionally good at processing large volumes of information, identifying statistical patterns and generating variations of possible ideas. Humans, by contrast, tend to be better at interpreting context, exercising judgement and recognizing when something does not quite make sense.
When the two are combined thoughtfully, the results can be powerful. Human creativity can be amplified by computational exploration, while machine-generated outputs can be filtered through human judgement. But the success of such a system depends on the relative strength of its components. Collective intelligence works best when the participants bring complementary strengths to the process.
If humans become overly dependent on machine-generated outputs, accepting them uncritically rather than interrogating them then the system becomes less intelligent rather than more. The technology begins to shape the thinking process rather than supporting it. This is sometimes described as ‘cognitive offloading’. Humans have always used tools to reduce mental effort. Writing allowed information to be stored outside the brain. Calculators reduced the need for mental arithmetic. Search engines transformed the way we retrieve information. Generative AI extends that trend into areas that were once central to knowledge work: constructing arguments, synthesizing information and articulating ideas. Tasks that previously required sustained cognitive effort can now be delegated, at least partially, to machines.
In many respects this is enormously useful. Few people would willingly abandon tools that accelerate research, streamline writing or open up new avenues for exploration. Yet there is an accompanying risk that deserves attention. When tools become sufficiently powerful and convenient, it becomes tempting to accept their outputs without subjecting them to careful scrutiny. The act of thinking gradually shifts from generating ideas to selecting between machine-generated alternatives. That shift may seem subtle, but over time it has implications for the development of judgement. And judgement is precisely the capability that philosophy has traditionally sought to cultivate.
For centuries philosophy has concerned itself with the discipline of thinking well. Philosophical training emphasizes the careful examination of assumptions, the identification of contradictions and the evaluation of arguments according to logical and ethical principles. Students of philosophy spend a great deal of time asking what appears, at first glance, to be an irritatingly simple question: how do we know that this claim is actually true? In many ways this tradition echoes the famous declaration of René Descartes, “I think, therefore I am,” which placed the act of thinking at the center of human existence.
The value of that discipline becomes clearer in a world where persuasive language can be generated at scale. Large language models are remarkably good at producing explanations that sound convincing, even when the underlying reasoning is incomplete or flawed. Their outputs are fluent, coherent and often persuasive. But fluency is not the same thing as truth.
Without the ability to interrogate arguments, examine assumptions and test reasoning, it becomes dangerously easy to mistake a well-constructed sentence for a well-founded conclusion. This is where philosophical habits of thought become increasingly valuable. They encourage intellectual skepticism, a willingness to question apparently obvious claims and a recognition that persuasive language can mask weak reasoning.
In that sense, philosophy can be understood not merely as an academic discipline but as a form of cognitive training. It strengthens the mental capacities that allow individuals to evaluate ideas rather than simply absorb them. As Socrates famously argued, “the unexamined life is not worth living.”
Artificial intelligence raises another set of philosophical questions that are equally significant. Throughout history, new technologies have repeatedly forced societies to reconsider moral boundaries and responsibilities. The industrial revolution reshaped labor markets and social structures. Nuclear technology confronted humanity with the destructive potential of scientific progress (“Now I am become Death, the destroyer of worlds” J. Robert Oppenheimer).
Artificial intelligence presents a similar challenge, although its effects unfold in different domains. The question confronting society is rarely whether something can be built. Engineers and entrepreneurs have demonstrated remarkable skill in pushing the boundaries of what is technically possible. The more difficult question is whether something should be built, and under what conditions.
Philosophical traditions provide frameworks for thinking about such dilemmas. Some approaches emphasize consequences and seek to maximize overall welfare. Others focus on principles, arguing that certain actions are inherently right or wrong regardless of the outcome. Still others examine the character and motivations of decision-makers themselves. None of these frameworks offers definitive answers. But they provide ways of thinking systematically about ethical trade-offs that might otherwise be addressed only after the fact.
It is therefore perhaps unsurprising that technology companies are beginning to draw on philosophical expertise as they navigate these questions. Yet the return of philosophy is not confined to AI firms. It is also quietly re-emerging as a practical skill within knowledge work itself. If artificial intelligence increasingly participates in the process of thinking, the value of human judgement rises rather than falls. Machines can generate possibilities at extraordinary speed, but they remain dependent on human interpretation to determine which possibilities are meaningful, relevant or ethically acceptable.
The human role in the system becomes one of evaluation rather than generation. And evaluation requires judgement.
This observation sits at the heart of the work we have been developing through PolymathMind. The framework emerged from a simple question: if machines become progressively better at producing information and performing analytical tasks, what capabilities remain distinctly human? The answer, we believe, lies in the strengthening of judgement, creativity and communication.
Our program therefore focuses on three broad episodes. The first concentrates on sharpening thinking — helping individuals clarify assumptions, interrogate arguments and recognize where human judgement remains essential. The second focuses on creativity, exploring how human imagination and contextual understanding can be combined with AI tools to generate something which is genuinely new. The third addresses communication, recognizing that persuasive narratives still depend on human understanding of meaning, emotion and audience. Seen from this perspective, PolymathMind is not primarily about technology. It is about strengthening the human side of an emerging human–AI cognitive system.
Which brings us back, in a rather roundabout way, to the philosophers now working inside AI firms. For much of the twentieth century philosophy was often treated as an abstract academic pursuit, intellectually interesting perhaps, but of limited practical relevance to the everyday world of business and technology. The curious twist of the AI era is that philosophy may be becoming practical again. Maybe educational establishments of the future will be led by a Head of Thinking with a PhD in classic philosophy & AI powered thinking tools.
As machines grow more capable at generating information and exploring analytical possibilities, the distinctly human capacities of judgement, reflection and meaning-making become more valuable rather than less. The future will not belong simply to those who build intelligent machines. It will belong to those who understand how humans and machines should think together.
As Søren Kierkegaard once observed, “life can only be understood backwards; but it must be lived forwards.” The same may prove true of the AI revolution.