Are humans unique, or are we just as deterministic and replicable as AI algorithms? In other words, do we have free will? These questions have challenged our sense of identity since Epicurus and Democritus debated in antiquity whether the world is governed by necessity and causal chains. But if AI, by contrast, can replicate and surpass us, could it also be equipped with empathy and sociality? The concept of determinism is relevant when contemplating an AI-driven future.
This is the third post in a trilogy about how AI can shape our perceptions of reality. This post addresses the fundamental question of what is real. Do humans, for instance, have free will, or is reality fundamentally deterministic? The first post explored whether AI can get to know us so well that it hyper-personalizes its communication, and if so, whether it can also predict and manipulate us. The second post discussed how AI often learns from data that it has generated itself.
The world has deterministic traits…
The world behaves deterministically at the level at which we operate. The universe, for example, can be described quite accurately through mathematics, such as the orbits of planets and the laws of nature.
Events and structures also tend to repeat themselves. On financial markets, for example, we see regular boom/bust cycles, and Fibonacci sequences have periodically strong predictive value.
This suggests that the world is fundamentally deterministic. The big question, then, is how we account for the moments when we cannot predict outcomes. Is this due to free will, or rather because of underlying structures we don’t yet understand or have language for?
… but lacks apparent deterministic lawfulness…
Since the early 20th century, we've learned that much in practice resists predictability:
- Certain quantum events are fundamentally unpredictable and can only be described probabilistically, such as the location of electrons. This is an expression of fundamental uncertainty, not merely technical inaccuracy.
- Chaos theory shows that even microscopic differences in initial conditions can lead to vastly different outcomes (e.g., Lorenz). This holds true even in deterministic systems like the weather, meaning that even though the system is deterministic, it’s practically unpredictable.
... including logically
In 1931, Austrian logician Kurt Gödel presented two incompleteness theorems. These state that there will always be statements in mathematics that are undecidable within a given formal system. They cannot, for example, contain internal proofs of their own consistency. There are thus fundamental limits to how much can be proven through formal methods. In other words, a fully deterministic system cannot contain all truths, cf. Alan Turing’s “halting problem.”
So does Free Will Exist ...
From an identity perspective, it is thus relevant to ask whether consciousness can introduce non-deterministic elements. Do we have any genuine freedom of choice?
- We can never truly know whether something is entirely determined, we simply work with the best models we have.
- In principle, this allows for determinism and free will to coexist. Determinism deals with facts, while free will deals with our interpretation of those facts.
- It’s similar to following sheet music, only the experience creates the melody. That freedom lies in consciousness, not necessarily in the structure of the system. It’s like working in a language where some concepts can't even be conceived, because the language itself prevents it.
- When reading the sheet music of Erik Satie’s Gymnopédies or Gnossiennes, they may appear nearly meaningless and arrhythmic. But when hearing them, they emerge as calm, harmonious, almost meditative pieces.
… or are we blinded by what we can measure?
A weakness of modern science is our fascination with what we can measure and test. We're skeptical of what we can only sense vaguely. The more we suspect there is more beyond the measurable, the more we tend to focus solely on what we can measure.
We often think and conduct research within a given worldview where naturalness and causality are the conditions of validity. Thus, we tend to optimize within the worldview rather than challenge its framework.
Finally, we are often limited by a “simplicity bias”, the tendency to assign higher truth value to simple systems. The simpler and more logical, the more we assume it must be true, cf. Ockham’s Razor.
An open worldview requires awareness of existing frames…
The discussion of determinism is more than a technical or philosophical exercise. It has implications for how we see ourselves as humans, especially concepts like intelligence, intention, and empathy. It also shapes our understanding of what AI can become.
- If we believe the world is fundamentally mechanical and fully modelable, we will design technologies and societies on that basis.
- But if we accept that both the world and we ourselves contain elements beyond measurement, control, and logic, we open up to another approach.
… and it demands the courage to transcend them
For AI may be able to go beyond the human “simplicity bias,” thanks in part to its superior computational power. This could enable the development of new kinds of “senses,” since our cognition is partly limited by the range of our existing sensory systems.
Moreover, AI opens up new epistemological possibilities because it is a new kind of intelligence—one that extends beyond our own. AI is more than just faster human intelligence. Israeli historian Yuval Noah Harari has therefore described it as Alien Intelligence, rather than Artificial Intelligence. Something genuinely new must emerge, otherwise, it’s not truly intelligent.
Thus, determinism and free will do not necessarily exclude one another. The distinction may lie more in the ability to transcend systemic complexity. Perhaps “free will” is merely our current term for those moments when our reason and understanding fail to explain the choices we’ve made.