Free Will and Determinism: From Human Nature to Artificial Intelligence

I. Introduction

One of humanity's oldest philosophical puzzles is the debate over free will vs. determinism. Do our actions proceed from prior causes, and are we therefore enslaved? Or are we free to choose? What used to be a purely philosophical question has become a scientific one, with support from neuroscience, psychology, and artificial intelligence.

II. The Scientific Basis of Human Decision-Making

A. The Neurological Foundation

Decision making starts in the brain. Neuroscience research, on the other hand, has shown that decisions result from interactions among many areas of the brain, especially the prefrontal cortex (the planning and executive centre of the brain) and the limbic system (the system responsible for emotions and memories).

In experiments performed in the 1980’s, Benjamin Libet showed that activity of the brain associated with movement occurs 550 milliseconds before we are consciously aware of deciding to make the movement (Libet, 1985). This finding, which runs against our prevalent common sense notion of conscious decision-making, revealed that some of our decisions might start unconsciously before we catch wind of them.

More recent research using functional magnetic resonance imaging (fMRI) – a method of measuring brain activity by measuring changes in blood flow – has gone further to show that brain activity patterns can even predict simple decisions up to 10 seconds before a subject thinks they have made anything more than a conscious choice (Soon, Brass et al., 2008).

B. The Role of Consciousness

However, consciousness seems important for decision making. Research shows that conscious awareness allows us to:

1. Manage impacts of changes cited in the initiated actions

2. Produce complex sequences of behaviour

3. This information must be integrated from multiple sources.

4. Think long term

In fact, the human brain unconscious processes a large amount of information, but consciousness seems to be required to integrate this information and make complex decisions (Dehaene et al., 2014).

C. The Illusion of Control

A fascinating aspect of human decision making, though, is that we believe we can control things more than we really can. Ellen Langer demonstrated this using several experiments: People often feel they control events that aren't controlled (Langer, 1975). For example:

- Instead, people like choosing their own lottery numbers instead of randomly getting those.

- People throw dice harder when they need high numbers

- Subjects think that they have the power to control random outcomes through practise.

This "illusion of control" appears to be a fundamental aspect of human psychology, possibly serving an evolutionary adaptive function by promoting active engagement with our environment.

III. Modern Scientific Perspectives

A. The Cognitive Science View

Modern cognitive science tells us that our decision making process doesn't happen purely deterministic or completely free; it's a framework of constrained flexibility. Research by Daniel Kahneman, recipient of the Nobel Prize in Economics, identifies two distinct systems of thinking (Kahneman, 2011):

1. System 1: Automatic, fast, and intuitive.

- Operates unconsciously

- Makes rapid judgments

- Prone to systematic biases

2. System 2: Slow, deliberative, and analytical

- Requires conscious effort

- Handles complex decisions

- System 1 impulse can be overridden

These systems operate together, with System 2 within reach of System 1 when required to help prevent automatic responses from occurring.

B. The Role of Uncertainty

Fundamental unpredictability in the physical world is introduced by Werner Heisenberg's uncertainty principle of quantum mechanics (Heisenberg, 1927). Some scientists have emerged with the idea that quantum effects in the brain could make room for free will. But, as physicist Max Tegmark wrote, most of what the brain does happens at scales far too big for quantum effects to matter (Tegmark, 2000).

IV. Experimental Evidence

A. Decision-Making Studies

Research on human decision-making has revealed several important patterns:

1. Unconscious Influences

Antonio Damasio has demonstrated that patients who have destroyed the centres that process emotions make bad decisions despite having normal logical reasoning. So, suggesting that emotions play a crucial role in decision making (Damasio, 1994).

2. Choice Architecture

Research by Thaler and Sunstein points out that the presentation of options itself changes our options, suggesting that free will operates in given environmental constraints (Thaler & Sunstein, 2021).

B. The Neuroscience of Free Will

Recent neuroscientific research has provided new insights:

1. As shown by Patrick Haggard's work (2008), our sense of agency ― that we’re responsible for our actions ― is built by the brain after the action occurs. So, therefore, that suggests that our experience of free will might be a useful biological construct and not a necessary property of decision making itself.

2. Using transcranial magnetic stimulation (TMS), a method that temporarily obstructs brain activity in specific areas, such studies have found that disrupting particular areas of the brain affects how we decide on something and whether we think we have freely chosen to do so (Brass & Haggard, 2007).

V. Quantum Mechanics and Free Will

A. The Copenhagen Interpretation

The Copenhagen interpretation of quantum mechanics, inspired especially by Niels Bohr and Werner Heisenberg, injects into physical reality a fundamental indeterminacy. While this doesn't prove the existence of free will, it demonstrates that the universe isn't strictly deterministic at its most fundamental level (Bohr, 1928).

B. Quantum Effects in Biology

Recent research in quantum biology has revealed that quantum effects can last much longer than has been assumed (Lambert et al., 2012). Of course, this does not necessarily endorse quantum theories of consciousness or free will; it does instead hint at the possibility that quantum effects may matter less grossly but more subtly in biological processes.

VI. Artificial Intelligence and Decision Making

A. Machine Learning Algorithms and Determinism

The way a modern artificial intelligence system, known as deep learning models, makes decisions is also quite similar to how we humans make decisions. These systems make decisions based on:

1. Training Data

Just like humans, neural networks are taught to recognise patterns in existing data from experience. Nevertheless, unlike humans, we can actually trace where these decisions are made. Yoshua Bengio shows through research (Bengio, 2009) that deep learning systems can act in seemingly autonomous ways while still being completely deterministic.

2. Probabilistic Decision Making

The decision making of most modern AI systems is a probabilistic behavior. ?For example, AlphaGo's victories over human Go champions weren't based on strictly deterministic calculations, but rather on probability distributions of potentially successful moves (Silver et al., 2016).

B. The Emergence of Complexity

Even though processes in complex AI systems are technically deterministic, their complexity makes behaviours impossible to predict practically. Emergent behaviour, as the name suggests, is similar to human decision making. Stuart Russell demonstrates that these seemingly autonomous behaviours are the result of collective, simple rules (Russell and Norvig, 2019).

VII. Ethical and Philosophical Implications

A. Moral Responsibility

The question of moral responsibility becomes particularly relevant as AI systems become more autonomous. Research by Peter Singer and others suggests that moral responsibility might not require libertarian free will (the ability to have done otherwise under exactly the same circumstances), but rather depends on:

1. The power to answer reasons.

2. Understanding consequences.

3. The ability to change behaviour when presented with new information (Singer, 2011)

B. AI Decision-Making and Accountability

Virginia Dignum’s recent work on the responsible use of AI examined how we can be sure that the decisions that AI systems make are in line with human values and accountable for those decisions. This involves:

1. Transparency in decision making processes

2. Explicability of AI actions

3. Deployment of AI responsibility frameworks (Dignum, 2019)

VIII. Future Perspectives

A. Human and Machine Decision Making Integration

As AI systems become more smart, David Chalmers has cited research which suggests that it will be harder to differentiate human and machine decision making (Chalmers, 2010). This raises important questions about:

1. Human-AI Collaboration

Studies published recently demonstrate that human-AI teams can perform better than either humans or AI on their own in lots of the domains (Brynjolfsson & McAfee, 2017).

2. Augmented Decision Making

The future may contain technologies that shore up human decision making capabilities, which could lead to some shifting of how we think about free will and determinism.

B. Implications for Society

The integration of AI into decision-making processes has broad societal implications:

1. Legal Systems

Responsibility and liability questions arising when AI systems make or help out make decisions.

2. Education

The urgency to create new ways of understanding and teaching decision making in an AI augmented world.

3. Ethics

New ethical frameworks that will allow us how both human and artificial decision makers.

Conclusion: The Dance of Consciousness and Decision Making

It turns out the relationship between consciousness, free will, and decision making is more complicated than you might have guessed. While Libet's experiments (1985) and subsequent research by Soon et al. (2008) suggest that decisions may be initiated before conscious awareness, this doesn't necessarily negate the role of consciousness in decision making. It suggests, however, a more complex interaction between conscious and unconscious processes.

The Learning Decision Framework

Think of learning to walk. Each step at first needs to be conscious and paying attention. This becomes automated over time, and we’re able to walk while having complex conscious thoughts. This is a pattern we see over and over within human behaviour; all you do is learn something consciously, and then you execute it subconsciously.

Important here is Damasio's (1994) research on the somatic marker hypothesis. Emotional experiences create biological markers that form future decisions, he found. These markers allow for rapid, seemingly automatic responses to similar situations, what he calls "gut feelings."

Evolutionary Perspective

From an evolutionary standpoint, this two-phase decision-making process makes perfect sense:

1. Conscious Learning Phase

- Detailed analysis of new situations.

- Outcomes considered carefully

- Response patterns

- Creation of emotional markers

2. Automated Response Phase

- Similar situation and a quick reaction

- Usage of cognitive resources

- Enhanced survival chances

This model explains why we can all make split second decisions in a dangerous situation and yet still be able to consciously assess new situations.

The Role of Consciousness in Decision Evolution

It's not that consciousness doesn't matter, but that maybe it isn't the final word when it comes to what you do. Perhaps it’s the constructor of our process for making choices. As has been pointed out by, for example, Kahneman (2011), System 2 (conscious) thinking can train and modify System 1 (automatic) "reactions" over time.

Joseph LeDoux's research (2019) further supports this view, showing how conscious processing can modify emotional responses and behavior patterns through a process he calls "synaptic plasticity" - the brain's ability to strengthen or weaken connections based on experience.

Implications for Free Will

This understanding requires a more sophisticated form of free will than the standard tradition of free will takes it to be. Rather than making each decision freely in the moment, we exercise our freedom through:

1. Complete consciousness in the learning and analysis of experiences

2. Response pattern development

3. Periodic review and modification of these patterns.

4. Conscious intervention beyond what we can do automatically

However, from this perspective, it seems that however mechanistic individual decisions might seem, it is the conscious mind which shapes the decision making architecture itself. Perhaps free will is to be better understood as the shaping of our pattern of decision making over time rather than moment to moment control. It reconciles the seeming paradox that our decisions can be predetermined and at the same time help us perceive agency and provides an explanation for how we can both learn by experience and respond instantaneously in an emergency.

Ultimately though, could it be from free will and determinism being mutually exclusive, or could they be two aspects views of a much richer reality in which consciousness shapes the very structure that later determines our choices?

Boldly go where no human or AI has gone before!

Ferhat Sarikaya

https://doi.org/10.5281/zenodo.13988196

References

[1] Bengio, Y. (2009). Learning deep architectures for AI. Foundations and Trends? in Machine Learning, 2(1), 1–127. https://doi.org/10.1561/2200000006

[2] Bohr, N. (1928). The quantum postulate and the recent development of atomic theory1. Nature, 121(3050), 580–590. https://doi.org/10.1038/121580a0

[3] Brass, M., & Haggard, P. (2007). To do or not to do: the neural signature of Self-Control. Journal of Neuroscience, 27(34), 9141–9145. https://doi.org/10.1523/jneurosci.0924-07.2007

[4] Brynjolfsson, E., & McAfee, A. (2017). The business of artificial intelligence. Harvard Business Review, 1–20. https://starlab-alliance.com/wp-content/uploads/2017/09/The-Business-of-Artificial-Intelligence.pdf

[5] Chalmers, D. J. (2010). The singularity: A philosophical analysis. The Australian National University. https://researchportalplus.anu.edu.au/en/publications/the-singularity-a-philosophical-analysis

[6] Damasio, A. R. (1994). Descartes’ error: Emotion, Reason, and the Human Brain. Putnam Adult.

[7] Dehaene, S., Charles, L., King, J., & Marti, S. (2014). Toward a computational theory of conscious processing. Current Opinion in Neurobiology, 25, 76–84. https://doi.org/10.1016/j.conb.2013.12.005

[8] Dignum, V. (2019). Responsible artificial intelligence: How to Develop and Use AI in a Responsible Way. Springer.

[9] Haggard, P. (2008). Human volition: towards a neuroscience of will. Nature Reviews. Neuroscience, 9(12), 934–946. https://doi.org/10.1038/nrn2497

[10] Heisenberg, W. (1927). über den anschaulichen Inhalt der quantentheoretischen Kinematik und Mechanik. The European Physical Journal A, 43(3–4), 172–198. https://doi.org/10.1007/bf01397280

[11] Kahneman, D. (2011). Thinking, fast and slow. Macmillan.

[12] Lambert, N., Chen, Y., Cheng, Y., Li, C., Chen, G., & Nori, F. (2012). Quantum biology. Nature Physics, 9(1), 10–18. https://doi.org/10.1038/nphys2474

[13] Langer, E. J. (1975). The illusion of control. Journal of Personality and Social Psychology, 32(2), 311–328. https://doi.org/10.1037/0022-3514.32.2.311

[14] LeDoux, J. (2019). The deep history of ourselves: The Four-Billion-Year Story of How We Got Conscious Brains. Penguin.

[15] Libet, B. (1985). Unconscious cerebral initiative and the role of conscious will in voluntary action. Behavioral and Brain Sciences, 8(4), 529–539. https://doi.org/10.1017/s0140525x00044903

[16] Russell, S., & Norvig, P. (2019). Artificial intelligence: A Modern Approach. Pearson Higher Education.

[17] Silver, D., Huang, A., Maddison, C. J., Guez, A., Sifre, L., Van Den Driessche, G., Schrittwieser, J., Antonoglou, I., Panneershelvam, V., Lanctot, M., Dieleman, S., Grewe, D., Nham, J., Kalchbrenner, N., Sutskever, I., Lillicrap, T., Leach, M., Kavukcuoglu, K., Graepel, T., & Hassabis, D. (2016). Mastering the game of Go with deep neural networks and tree search. Nature, 529(7587), 484–489. https://doi.org/10.1038/nature16961

[18] Singer, P. (2011). Practical ethics. https://doi.org/10.1017/cbo9780511975950

[19] Soon, C. S., Brass, M., Heinze, H., & Haynes, J. (2008). Unconscious determinants of free decisions in the human brain. Nature Neuroscience, 11(5), 543–545. https://doi.org/10.1038/nn.2112

[20] Tegmark, M. (2000). Importance of quantum decoherence in brain processes. Physical Review. E, Statistical Physics, Plasmas, Fluids, and Related Interdisciplinary Topics, 61(4), 4194–4206. https://doi.org/10.1103/physreve.61.4194 ?

[21] Thaler, R. H., & Sunstein, C. R. (2021). Nudge: The Final Edition. Yale University Press.

?