The Power of Human Heuristics

The term ‘creativity’ often summons images of grand artistic achievements: a novelist crafting a timeless story, a musician composing an unexpected symphony, or an artist whose canvas redefines a genre. 

These are the triumphs of exceptional minds, individuals who transcend the ordinary to produce works of singular brilliance. Yet creativity is not the preserve of such rare talents. It flourishes quietly in the everyday, in moments when ordinary people confront life’s challenges. This is the creativity of problem-solving—the ability to navigate a multitude of possibilities and select a path that, while not always perfect, is functional. It is the art of discarding dead ends and finding clarity amid complexity. This understated, pervasive ingenuity, often overlooked, forms the bedrock on which the future of creativity will be built.

This issue is not merely philosophical but also technical. Recently, robotaxis—autonomous vehicles developed by Elon Musk’s company—began trials in Austin. These cars, guided by artificial intelligence algorithms rather than human drivers, revealed significant flaws. During testing, they dropped off passengers in unsafe locations, such as the middle of multi-lane roads or at busy traffic-light junctions. Even a novice human driver would avoid such obvious errors. How could sophisticated AI systems falter so glaringly?

Heuristics as the Key

The answer lies in ‘heuristics’—experience-based approaches that deliver ‘good enough’ solutions quickly, with minimal time and energy having been expended. Unlike optimal solutions, heuristics rely on estimation, intuition, experience, or common sense. They are not guaranteed to be perfect but are versatile, straightforward, and rapid, allowing for refinements over time.

Artificial intelligence strives for optimal outcomes, whether in generating text, designing machine components, or planning logistical routes. Its strength lies in its combinatorial capacity—the ability to sift through vast arrays of possibilities and apply known solutions. Yet this brute computational power is insufficient on its own. AI also depends on heuristics—rules that eliminate unpromising options to streamline decision-making. For instance, when buying a car, ‘red flags’ such as discrepancies in the service history, missing oil change records, or high mileage on rough roads signal potential issues. These heuristics, often rooted in deep personal experience, are tied to an individual’s history and values. Many are unformalisable, known in AI terminology as ‘tacit knowledge,’ and humans can combine them instinctively to make swift judgements. The best heuristics reflect a person’s unique needs, interests, and worldview.

The Energy Efficiency of Human Creativity

Heuristics are energy-efficient. They can appear as biases or overly conservative assumptions, which may become outdated and prevent better solutions from being found. Yet their primary function is to save time and effort. In high-stakes situations—braking to avoid a collision, redirecting a plan, or selling a stock—humans must act quickly, relying on the limited energy derived from food, not megawatts from an electrical grid.

The human brain, consuming just 20 watts—equivalent to a standard light bulb—manages complex tasks like decision-making, planning, and creation. It achieves this not by exhaustively analysing every possibility but by instantly ruling out most. Heuristics—unspoken, experience-based rules—enable the brain to label something as ‘nonsense’ before fully articulating the problem. By contrast, AI may burn through megawatt-hours to reach a similar conclusion, if it does at all. This ability to eliminate dead ends—rapidly, intuitively, and with minimal energy—gives humans a distinct advantage.

AI as a Derivative of Human Intelligence

AI’s learning process is not autonomous; it is an infrastructure built on human decisions, values, and heuristics. The selection of data, model architectures, and performance metrics is a product of human ingenuity. A model may process trillions of words, but its training reflects human choices—what to include, what to exclude, what constitutes an error, and what defines a useful pattern. In this sense, AI is a powerful derivative of human intelligence, not its replacement. It functions more like a calculator than an artist or scientist. It does not question what is worth exploring, know when to stop, or intuit when a solution is absurd because it lacks the concept of absurdity. AI is devoid of experience, values, and intuition—the very foundations of true creativity.

While AI may surpass humans in speed, precision, and computational power for many tasks, human creativity excels in an energy-constrained world through the often-unconscious ability to dismiss fruitless paths. This imperfect, value-driven efficiency makes human thinking profoundly economical and genuinely creative. Without these deeply human heuristics, AI risks floundering in an ocean of infinite possibilities, rarely uncovering anything truly valuable without immense energy costs. 

It is human heuristics, born of lived experience rather than machine learning or algorithms, that make AI useful. The creativity of tomorrow will hinge on blending humanity’s intuitive grasp of what matters with AI’s computational strength, which in turn would amplify and harness that intuition. If AI’s future lies in mimicking human judgement, the next frontier is not more data but better encoding of human heuristics. Without them, intelligence may remain just artificial, and not creative.

Statement

Creativity isn’t just for artists—it’s the everyday genius of solving life’s puzzles with intuition and experience. While AI’s raw computing power stumbles, dropping off robotaxi passengers in dangerous locations, human heuristics swiftly cut through complexity with minimal energy. These unspoken rules, rooted in values and tacit knowledge, make human thinking uniquely efficient. As AI leans on human ingenuity, the future of creativity lies in marrying our intuitive spark with computational might.