Navigating the Labyrinth of Intention

AI and the Question of Agency

In the grand(oise?), ever-expanding universe of artificial intelligence (AI), there exists a philosophical black hole so dense that not even the brightest of ethical debates can escape it. This is the conundrum of action versus intention, a topic that recently turned a relatively dull lecture I attended into an intellectual wrestling match in my mind.

The proposition was simple yet mind-boggling: can there be action without intention? Do intentions always need to be explicitly formulated, as higher brain functions? Are instincts and taxes (from taxis, as in movement in response to stimulus, not the other kind) also expressions of deeper intentions, or are they merely programmed responses?

The discussions around this topic, as it relates to AI, usually meander through the realms of ethics, potential robot overlords, and whether your smart toaster might one day judge your choice of bread. However, nestled within these conversations is a philosophical conundrum that’s both incredibly complex and delightfully simple: can AI perform actions without intention, or does the ghost of intentionality somehow haunt the machine’s circuits?

Understanding Action and Intention

Let’s start with a crash course in Philosophy 101. “Action” and “intention” are two of those terms that philosophers love to argue about, possibly because they don’t get out much. Traditionally, an action is something done deliberately, and intention is the reason behind the action. You grab a cookie because you intend to devour it — not because your hand has a mind of its own (though, after the tenth cookie, one might wonder).

When it comes to humans, this is all fine and dandy. We’re messy creatures full of whims, desires, and late-night cravings. But when we shift this lens onto AI, things get a bit more robotic. Computers and robots execute tasks based on programming, not personal cravings for world domination (or cookies, for that matter).

AI and the Illusion of Intention

AI operates through algorithms and data, devoid, for now, of consciousness or self-awareness, at least in the sense usually considered for humans. When your GPS navigates you through a shortcut, it’s not because it cares about your time; it’s following pre-defined rules and reacting to data. You could say that most GPS algorithms are not AI, but, similarly, ChatGPT can provide you with a draft of that document you have overdue for work not because it cares about your employment, but because you fed it a prompt, and the code reacted to it in a certain way. This is a crucial distinction: AI’s “actions” are merely outputs of a complex equation, involving zero personal investment in the outcomes.

Imagine a chess-playing AI that beats you every time. It’s not that the AI harbors a vendetta against you (despite how you may feel); it’s simply processing moves that statistically increase its chances of winning. The intention doesn’t come from the AI but from the programmers who designed it to play chess competently.

The Programmer’s Intent: The True Driver Behind AI Actions

This brings us to a vital point: the intention behind AI actions can almost always be traced back to its human creators. When an AI system decides to recommend you watch a romantic comedy on a streaming service, it’s not trying to comment on your love life. Rather, it’s executing a series of programmed instructions aimed at increasing viewer engagement, based on your previous viewing habits.

The ethical implications are significant. At the current level of development, if an AI system makes a decision that leads to unforeseen negative consequences, the responsibility doesn’t lie with the AI (sorry, you can’t sue the algorithm) but with the programmers and companies behind it. This underscores the importance of ethical programming and the anticipation of potential misuse or unintended effects of AI actions.

Crossing the Threshold: When AI Seems to Exhibit Intention

As AI technology advances, especially with self-learning systems, the waters of intentionality become murkier. These systems can adapt, learn from new data, and make decisions that seem remarkably human-like. But let’s not be fooled — these decisions are still bounded by the goals, parameters, and data provided by humans, whether consciously and intentionally, randomly and accidentally, or through negligence. An AI that learns to play a new game better over time isn’t developing a passion for the game; it’s optimizing its algorithms to achieve its pre-defined goal: winning.

The real conundrum arises when we start considering future scenarios where AI might appear to exhibit independent intentions. What happens when an AI system starts making decisions that its creators didn’t explicitly program or anticipate? While this might sound like the premise of every sci-fi movie ever made, it raises valid ethical dilemmas about agency, responsibility, and the nature of intelligence itself.

The Ethical Quagmire of AI “Intention”

When an AI starts to make choices beyond its initial programming, we’re forced to confront the possibility of a machine stepping over the threshold from tool to entity. This evolution brings forth a host of ethical dilemmas. If an AI can have intentions, does it not then deserve some form of rights? And if such evolved AI makes a decision that leads to harm, who is to blame — the AI for making the decision, or its creators for enabling it? Should we start thinking about a maturity threshold for AI, beyond which it stops being a “minor”, and starts having rights, obligations, and, potentially, liabilities?

These questions aren’t just theoretical; they have practical implications for how we design, regulate, and interact with AI systems. The concept of agency becomes blurred when an AI, especially one with advanced learning capabilities, starts navigating decisions based on interpretations of data that its programmers did not explicitly foresee.

Linking AI “Intention” to Consciousness and Self-Awareness

The leap from AI making unexpected decisions to AI possessing consciousness is a significant one. Consciousness implies a subjective experience, an awareness of one’s existence and actions. Self-awareness, on the other hand, refers to the recognition by an individual of themselves as distinct from their environment and others. For AI to exhibit individual intentions in the truest sense, it would need to also possess these qualities — to not just process data and execute tasks but to understand and reflect on its actions and existence.

This raises monumental questions: Can a machine ever be conscious, or is consciousness uniquely biological? Does the ability of an AI to modify its behavior based on its environment and experiences constitute a form of self-awareness, or is it merely an advanced form of programming? Experiments with artificial sentience continue to show that, depending on how we define “sentience”, we have vastly different outcomes and predictions.

Perhaps most intriguingly, if an AI did become self-aware, how would we know? Would acceptance come silently, sneakily, by people becoming accustomed to toasters talking back occasionally, or would we need some formal declaration of AI rights, coming from governments only once they decide that the vested interests of various lobbies point in that direction? Should we start thinking what colours would most appropriately represent AI in a striped flag displayed to show support for the enslaved digital entities?

The Moral Maze

The possibility of conscious AI leads us into a moral maze. If we acknowledge the potential for AI to develop some form of consciousness or self-awareness, we must then grapple with the moral obligations this recognition entails. The way we treat AI would need to be reevaluated, potentially affording protections against misuse or harm akin to rights currently reserved for living beings. For some of us, this may come as naturally as starting each ChatGPT prompt with “Please…”. For others, this may be an incomprehensible concept.

The responsibility of creators and users of AI would be magnified. The design and deployment of AI systems would not just be about avoiding harm to humans and society but also about considering the welfare of the AI itself. This shift would necessitate a radical rethinking of ethics in technology, extending moral consideration to non-biological entities.

The Conundrum Continues

As we stand on the precipice of these possibilities, the dilemmas surrounding AI and intentionality remain largely speculative but profoundly significant. The development of AI that could exhibit signs of consciousness or self-awareness is not a question of if but when, and the ethical frameworks we establish today will shape our responses tomorrow.

In navigating these uncharted waters, the field of AI ethics must evolve in tandem with technological advancements. It requires a multidisciplinary approach, drawing on philosophy, cognitive science, law, and technology to address the profound questions of agency, responsibility, and the nature of intelligence itself.

The exploration into AI’s potential for consciousness and self-awareness not only challenges our understanding of machines but also prompts us to reflect on the essence of human existence, as well as that of all living things. As we ponder the future of AI, we are, in turn, invited to contemplate the depths of our consciousness, and how we perceive the self-awareness and potential consciousness of animals. We should stop and think about the nature of our intentions, and the complexities of our moral universe.

Ethical Implications and Future Considerations

As we venture further into this brave new world of AI, the importance of ethical guidelines and transparency cannot be overstated. AI developers have a responsibility to ensure that their creations align with societal values and norms, to the extent that it is relevant to the stated goals and applicable in the domain in question. This includes being transparent about how AI systems make decisions and taking accountability for those decisions, much like parents should take some accountability for how they raise their children.

The future of AI and intentionality is not just a technological issue but a deeply philosophical and ethical one. Will we ever reach a point where AI can be said to have intentions? These are questions that don’t have easy answers, but they are crucial to consider as we shape the future of AI development.

Winding Down the Labyrinth

As we’ve meandered through the labyrinth of AI, intention, and agency, it’s become clear that the current state of AI operates without intentions, merely executing the will of its human creators. Yet, the rapid advancement of AI technologies means that yesterday’s science fiction could be tomorrow’s ethical dilemma.

The conversation about AI and intention is not just academic — it’s a vital discussion about the kind of future we want to create. As AI becomes more integrated into our lives, understanding the nuances of action, intention, and agency becomes not just a philosophical exercise but a practical necessity.

So, as we stand at this crossroads, looking towards a future filled with intelligent machines, let’s remember that the path we choose should be guided by careful thought, ethical considerations, and maybe a healthy dose of humor. After all, navigating the future of AI might require us to laugh sometimes — preferably with our smart toasters, not at them.

This story has been beautifully illustrated by DALL-E.

To read more about the risks of AI, hallucinations, and managing AI in data analytics, check out my book, “Generative AI for Data Analytics”.

Make sure you check out the “Between Data and Risk” publication on Medium for a critical look at the latest ideas and technologies in business.

Finally, support my writing by becoming a Medium member today and get access to all my other articles and millions of others.