Some years ago, I read an article by George Dvorsky, a Canadian expert in futurism, spaceflight, bioethics, and the search for extraterrestrial intelligence, that has stayed with me ever since.

In his article, "8 Great Philosophical Questions That We’ll Never Solve," published on Gizmodo, Dvorsky unpacks timeless questions that challenge how we think about existence, knowledge, and consciousness. One of those questions—about free will—profoundly inspired me and has shaped how I think about AI, decision-making, and leadership from those early days in my career in AI.

Later, that reflection resurfaced and was deepened by a course I took some years ago by the Massachusetts Institute of Technology, titled "Minds and Machines: Philosophy & Ethics." A rigorous 12-week learning path introduced me to the contemporary philosophy of mind, exploring consciousness, artificial intelligence, and the ever-blurring line between reality and perception. It asked tough questions about the relationship between the mind and the body, the nature of thought, and whether machines could ever genuinely think.

According to many scientists and philosophers, explaining consciousness is one of the deepest intellectual challenges of our time. That challenge continues to echo through my work in AI leadership. And from those studies, one main question still remains among my deepest reflections:

Do we have free will?

I often wonder whether our choices are truly our own or merely the result of a complex chain of causes we didn’t choose.

This age-old question is more than a late-night philosophical puzzle. It actually offers a powerful lens for understanding Artificial Intelligence, how we design it, and how we lead responsibly in an AI-powered world. And even if this topic may sound too philosophical, as an AI leader, I believe this matters more than we might think.

This question has captivated philosophers for centuries: are we truly free, or are we simply acting out a sequence of decisions set in motion by our biology, upbringing, and the world around us? The answers aren’t easy or singular. But over time, several major philosophical perspectives have emerged, each offering a different lens through which to explore what it means to choose freely.

Before we tie this to how AI makes decisions, it’s worth taking a brief look at the main ideas in the debate around free will:

• Determinism argues that everything we do stems from prior causes.
• Indeterminism counters with randomness, though chance alone doesn’t equate to meaningful freedom.
• Compatibilism offers a middle ground, suggesting that we are free as long as we act according to our reasons and desires, even if those reasons and desires are themselves determined.
• Then there’s the libertarian view, which insists we do have real agency and the ability to choose otherwise.

These ideas might sound too much abstract, but they gain practical weight when we think about how AI systems “make decisions.”

Traditional AI models, like rule-based systems, are clearly deterministic: feed them the same input and they will always yield the same output. Their behavior is predictable, fixed by the code and logic designed by humans.

More recent AI, especially machine learning and deep learning systems, appear less rigid. They include randomness in training or decision-making and produce outputs based on probabilities. These systems can behave in ways that seem creative, even surprising. Yet they still lack intention. They do not decide in the way humans do. They don’t ponder, reflect, or feel responsible. They calculate. This also includes your favorite Generative AI model or application as well.

But before we assume that all unpredictability hints at intelligence, we should consider how AI quietly influences everyday life—even in moments that feel personal and spontaneous. Sometimes, the systems we use aren’t just tools; they are framing the decisions we think we’re making.

For example, when my wife spends what feels like an eternity scrolling through Netflix to choose our next movie or series, she likes to believe it's entirely her decision. But is it really? Or has an algorithm already shaped that choice long before she made it?

I always try to remind her that Netflix doesn’t just offer a list of options—it curates them based on her watch history, preferences, behavior patterns, and comparisons with similar users. The illusion of choice is real: it is like when we see a menu, but the restaurant has already decided what’s on it, and maybe worse, already decided what we are going to eat...

This subtle influence, where algorithms shape outcomes without our realizing it, mirrors a broader pattern we observe in AI systems across various industries.

But probably, we are all fine when an algorithm helps us navigate blindfolded and choose the next series we're going to dive into during the weekend, right? However, let's consider a more serious and complex context, such as healthcare.

A neural network used for diagnostics might recommend a treatment not because it 'understands' the patient's emotional state or long-term goals, but because it has learned, from thousands of past cases, what’s statistically likely to be effective. There’s no empathy, no context, no intention—only probability.

This is eerily close to how indeterminism introduces chance into human decisions: a layer of unpredictability, but still without awareness. AI, like our philosophical dilemmas, shows us that complexity and choice are not the same as freedom. And that realization sets the stage for a deeper responsibility.

And this is where the reflection on free will becomes practical (and interesting!), because if AI doesn’t possess free will—and I guess we all accept that it doesn’t—it also doesn’t carry moral responsibility. For example, when an AI system recommends a job applicant, sets an insurance rate, or proposes a medical treatment, it’s not deciding with (human) intent. It is following patterns and probabilities derived from data. Basically, your AI system may be as sophisticated as you want, but it is ultimately a mirror. And what it reflects depends entirely on us, humans!

I know it may sound a little bit uncomfortable, but this realization exposes the danger of what has been called the “responsibility gap.”

When something goes wrong in an AI-driven process, such as bias, discrimination, or harm, we, humans, are tempted to treat the system like a black box or a neutral agent. But (fortunately) algorithms don’t exist in a vacuum. We, the people, design them. And we, the people, choose the data. We, the people, decide how and where to apply them.

Remember: behind every AI decision is a chain of human decisions, whether thoughtful or careless.

Let me share an example: do you remember the now-shelved AI recruiting tool developed by Amazon? The model was trained on historical hiring data, which was heavily skewed toward male candidates. The result? It started downgrading CVs with indicators of being female, like references to women’s colleges. The AI didn’t intend to discriminate; it simply optimized based on biased data. But we, the people, know that it was up to Amazon to recognize and stop it, right? This isn’t just an AI error—it’s a human accountability issue.

That said, unfortunately, I need to remind us all that the ethical center of AI lies with us. Leaders, developers, policymakers—we are the ones who remain accountable. And this accountability is not something we can outsource to the machines, no matter how advanced they become.

If we want our AI systems to be fair, transparent, and aligned with our values, we need to build those values in from the beginning, by consciously addressing bias in the training data. It means creating mechanisms for transparency, so people understand how and why decisions are made. It means maintaining oversight, ensuring there is always a human who can intervene when necessary.

But, backing to the concept of free will, in the age of AI, it must become more than a philosophical curiosity. It’s becoming a call to responsibility.

If our machines are deterministic or probabilistic systems with no agency of their own, it is up to us to lead with ethical clarity. We must be more vigilant than ever about how we design and deploy these systems, especially when their outputs impact lives and livelihoods.

Of course, I am not suggesting that your CEO must evolve into some sort of philosopher 2.0 solely because of AI, but I am calling for reflection. This new era of hyper-automation requires all of us to stay curious about how decisions are made, not only by our machines but also by us, humans.

This AI era reminds us that our tools reflect our intentions. And it reinforces a truth we often overlook: we are still the ones deciding. Not the AI. Not yet.... not yet!

In a world increasingly shaped by algorithms, remembering that distinction may be one of the most important decisions we ever make.

But there’s more to this conversation. Let’s go deeper.

As AI becomes more embedded in our decision-making processes, we encounter a critical limitation: the illusion of explainability.

In philosophy, particularly in debates around consciousness and knowledge, there’s an acknowledgment that true understanding isn’t just about outcomes—it’s about how those outcomes come to be. In AI, we often settle for knowing what the output is, but not why. This is especially true with deep learning models, which many of us label as the quintessence of the black boxes.

Deep Learning offers results—sometimes powerful, sometimes flawed—but lack transparency. And without transparency, how can we trust or challenge a system’s logic?

The ability to explain decisions, something even the most deterministic human can attempt, is largely absent in AI, at least by default. This is why ethical design must include explainability, or at least interpretability, at its core. Not just because it makes systems clearer, but because it makes responsibility possible.

There’s also a quieter risk: moral dilution. As more decisions are made—or appear to be made—by systems, the lines of responsibility can blur. When AI is in the loop, it’s easy to say “the system recommended it,” and subtly shift blame. But this diffusion of responsibility is a problem. It echoes a deeper philosophical concern: if everyone is responsible, no one really is. We need structures that keep human accountability visible, even when decisions are automated. Someone must be responsible for how the model was trained, validated, and monitored. Otherwise, ethics gets lost in the noise.

And finally, perhaps the most uncomfortable reflection: what if we are not as different from machines as we’d like to believe? The more we build systems that mirror human decision-making—language generation, pattern recognition, creative outputs—the more we confront the possibility that human choice might also be a kind of highly evolved computation.

Philosophers call this the challenge to human exceptionalism. In this view, the real difference is not in the act of choosing, but in the self-awareness that surrounds it. Humans reflect, feel, take responsibility. Machines don’t. Yet. This is where leadership becomes essential. If we believe in human agency, then it’s our job to act on it—to guide AI, govern it, and ensure it reflects not just what we can do, but what we should do.

So as we continue designing AI systems that assist, augment, or even challenge our decisions, we must keep returning to that core reflection: what does it mean to choose? And more urgently: who is doing the choosing now?

In a world increasingly shaped by algorithms, remembering that distinction may be one of the most important decisions we ever make.