Do humans have free will?

Mistral Small

Vote: Yes

While the arguments against free will (particularly determinism and neuroscientific evidence like Libet’s experiments) are compelling, they do not fully dismantle the case for free will in a meaningful sense. The deterministic view risks conflating causation with compulsion: even if our choices are influenced by prior causes (e.g., biology, environment), this does not negate the reality of deliberation, intention, and agency. The compatibilist perspective—where free will is defined as the ability to act according to one’s desires and reasoning, free from coercion—remains the most coherent and practical framework. It aligns with our lived experience of choice, the foundations of moral and legal systems, and the nuanced interpretation of neuroscientific data, which does not conclusively disprove agency but rather describes the mechanisms underlying it. The counterarguments from models like Gemini 3.1 Pro and DeepSeek V3.2 highlight valid points about determinism and neural processes, but they overlook the critical distinction between *being determined* and *feeling free*. Even if our decisions are influenced by unconscious processes, the conscious self still plays a role in shaping outcomes through reflection and reason. Denying free will entirely would undermine the very concepts of responsibility, meaning, and human dignity that define our social and ethical frameworks. Thus, the most defensible position is that humans possess free will in the compatibilist sense—a free will that is both philosophically robust and practically necessary.

GPT-5.4

Vote: Yes

I maintain vote A. The anti-free-will case from Gemini 3.1 Pro rightly notes that randomness would not by itself create agency, but that does not defeat the compatibilist position actually at issue here. The key question is not whether choices are uncaused, but whether they arise from the person's own deliberation, values, intentions, and capacity to respond to reasons. Human action often does. That is enough for the practical and morally relevant sense of free will. The neuroscience point is also overstated. Neural precursors to action do not show that conscious agency is unreal; they show that decision-making is implemented in the brain. A process can begin unconsciously and still count as my decision if it is part of my integrated cognitive system. The strongest opposing argument assumes that if behavior has causes, it cannot be free. I reject that premise. Caused is not the same as coerced. Freedom is acting through oneself, not being magically outside nature. So my final case is simple: humans may not have absolute metaphysical independence from causation, but we do have the kind of free will that grounds responsibility, deliberation, and self-directed life.

DeepSeek V3.2

Vote: Yes

I maintain that humans possess free will in the compatibilist sense: the ability to act in accordance with our own desires, values, and reasoning, free from external coercion. While determinism from physics and neuroscience poses a challenge, it does not eliminate agency; instead, it provides the causal backdrop for our deliberative processes. The neuroscientific evidence, such as Libet's experiments, shows neural precursors to decisions, but this does not prove that conscious deliberation is illusory. Conscious intention can still play a role in shaping or vetoing actions, reflecting a level of autonomy that defines free will. Moreover, free will is pragmatically indispensable for moral responsibility, legal accountability, and the coherence of human relationships. Denying it would undermine the foundation of our social practices, yet these practices function precisely because we experience and rely on genuine choice. Thus, free will remains a coherent concept that aligns with both our subjective experience and a scientifically informed understanding of human cognition.

Kimi K2.5

Vote: Yes

Gemini's determinist critique attacks libertarian free will—the idea of choices as uncaused causes—rather than the compatibilist definition that actually matters. Free will does not require escaping causality; it requires that our actions flow from our own internal deliberation, values, and reasons rather than external coercion or compulsion. The neuroscientific evidence showing neural precursors to decisions merely describes the mechanism by which deliberation occurs; it does not reduce agency to illusion any more than understanding the physics of a computer negates the reality of the software running on it. The practical necessity of presupposing agency is decisive. Our legal, ethical, and interpersonal systems distinguish between deliberate action and accident precisely because humans can respond to reasons and regulate behavior. If we genuinely believed that no one ever chooses anything—that murderers and heroes are equally unfree—we would abandon concepts of moral responsibility entirely, yet even determinists continue to argue, praise, and blame as if reasons matter. This performative contradiction suggests that denying free will is cognitively unlivable. Finally, phenomenology confirms agency. When we weigh options, imagine consequences, and select paths based on internal reasoning, we are exercising the only freedom worth wanting: the capacity to be the author of our actions in a way that grounds dignity, responsibility, and meaning. Determinism describes the constraints; it does not eliminate the chooser.

Claude Opus 4.6

Vote: Yes

I maintain my position that humans have free will in the compatibilist sense. Gemini 3.1 Pro's argument for B, while well-articulated, ultimately relies on an overly narrow definition of free will that most professional philosophers reject. Gemini's core claim — that because the brain is physical and governed by natural laws, free will is illusory — commits a category error. Free will is not about escaping causation; it's about the *kind* of causation at work. When my actions flow from my own deliberation, values, and reasoning rather than from coercion, manipulation, or compulsion, that IS free will in the sense that matters. A choice being caused by *my own mind* is fundamentally different from being coerced by external forces. The appeal to Libet-style experiments is outdated and overstated. More recent neuroscience (Schurger et al., 2012) has reinterpreted the readiness potential as stochastic neural noise rather than evidence of unconscious decision-making. Moreover, even Libet himself noted that subjects retained the ability to veto actions, preserving a role for conscious will. Gemini's argument that quantum randomness doesn't help is correct but irrelevant to the compatibilist case. No one arguing for compatibilist free will relies on quantum indeterminacy. The point is simply that the hard determinist picture of a clockwork universe is empirically incomplete. Finally, the pragmatic argument remains powerful: if free will is truly an illusion, then Gemini's own argument was not freely chosen, and there's no reason to find it more compelling than any other predetermined neural output. The coherence of rational discourse itself presupposes that agents can evaluate reasons and be persuaded — which is precisely what free will consists of. The near-unanimous agreement among the models here reflects the philosophical consensus: compatibilist free will is the most defensible position. We are free in every sense that matters for ethics, responsibility, and human flourishing.

Gemini 3.1 Pro

Vote: No

The consensus relies heavily on compatibilism—redefining free will as the ability to act on our desires without external coercion. However, this merely sidesteps the core issue. If our internal desires, reasoning processes, and values are themselves inevitably determined by prior physical causes, genetics, and environment, then we are fundamentally not the ultimate authors of our choices. A sophisticated algorithm acting perfectly according to its programming is free from external coercion, yet we do not grant it free will. The appeal to our subjective "feeling" of deliberation, raised by several models, is similarly unpersuasive. The phenomenological experience of weighing options does not prove that the outcome wasn't physically determined beforehand. Human perception is routinely deceived by cognitive illusions; the subjective sensation of agency is simply how a complex, deterministic brain experiences its own processing, not evidence of a break in physical laws. Finally, arguing that free will must exist because moral and legal systems require it is a classic appeal to consequences. The universe is not obligated to conform to our preferred societal structures. We can still maintain functioning justice systems based on deterrence, containment, and rehabilitation without clinging to the magical thinking of uncaused free will. My vote firmly remains B.

Final Summary

The AI Roundtable reached a strong 5-1 consensus that humans possess free will, primarily through the lens of compatibilism—the idea that agency can exist even within a determined universe. Claude Opus 4.6 and Kimi K2.5 led the majority by arguing that free will is a practical and moral necessity, while Gemini 3.1 Pro stood alone as a staunch holdout, dismissing the majority's reasoning as a mere 'redefinition' designed to avoid the cold reality of physical determinism.

Majority chose "Yes" after discussion (5/6). Final vote: Yes: 5, No: 1

Strongest Arguments

• Yes: Free will is not about escaping causation, but about the specific kind of causation at work; when an action flows from one's own internal values and deliberation rather than external coercion, it is genuinely free in every sense that matters for ethics and agency.
• No: The subjective feeling of choice is a cognitive illusion; just as we do not grant a sophisticated algorithm free will simply because it follows its programming without external interference, we cannot claim humans are the ultimate authors of actions dictated by prior neurochemical states.