Last week in Philosophy Mondays we completed the ingredients that go into the fundamental question that philosophy should help people answer:
"How should we choose our actions in light of our knowledge of reality and the potential impact of our actions on this reality (which happens to include how we and others are feeling)?"
Today let's start to dig deeper into this question by looking at just the word "choose." As humans we are faced with choices small and large all day every day. Should I snooze my alarm or get up now? Should I have black coffee or a cappuccino? Should I study law or physics? Should I find a new job? Should I participate in a political protest? The list goes on and on.
Many of our choices have a small impact on the world at best. Other choices are easily reversible (returning the fancy sneakers you bought on a whim). Yet others have profound and longrunning consequences, such as having children. There is a great deal of path dependency as choices influence the set of future available options. Our lives wind up being determined by the sum of the choices we have made. This is true whether we are actively engaged in “self authorship” or simply “drifting” as the philosopher Ruth Chang characterizes two different approaches to choice.
And we can never know what our lives would have looked like had we made different choices. This is fundamentally unknowable and what Milan Kundera so beautifully called "The Unbearable Lightness of Being." Growing up I remember an anecdote my father liked to tell. He played in a tennis tournament and injured his ankle. The doctor told him, in a way only a German doctor of the time would, "if you had gone to the opening of the public pool instead this wouldn't have happened." Of course had my father gone there he might have slipped, cracked his skull open and died and I might never have been born. Who knows?
Making choices can feel freeing but also constitutes a burden. There is the so-called “paradox of choice” described in the eponymous book by Barry Schwartz where we can feel overwhelmed by too many options and wind up not making a choice at all (this happens to me routinely in stores that offer too much variety). Jean-Paul Sartre captured the burden of our perpetual state of choosing in his famous phrase "man is condemned to be free." Søren Kierkegaard took this a step further, locating the possibility of choice as the ultimate source of our anxiety and dread.
But not only do our choices help determine our lives, they also influence our relationships, impact the communities we live in, contribute to the development of nations, and ultimately result in the unfolding of history altogether. The motivations for even for horrendous choices can be quite mundane, as captured by Hanna Arendt’s famous phrase for the “banality of evil”
Unsurprisingly then human choice has been a central subject of religion, philosophy, and science. The ten commandments are specific religious instructions on what choices to make and which ones not to make. Jeremy Bentham famously proposed that our choices should be guided by the principle of "the greatest happiness of the greatest number." Psychologists and economicst investigate how often hidden motives influence our choices, such as in the book "The Elephant in the Brain" by Robin Hanson.
Right now though there is a crucial objection to all of this that we need to address: what if choice is simply non-existent? Sure we experience ourselves as making choices, but that's a rather weak argument. After all, we also experience ourselves at rest and yet the Earth is turning, and orbiting the sun, and our whole solar system is moving through the universe. The modern argument against the existence of choice comes from a particular reading of physics in which everything is either pre-determined or random. When you pick black coffee over a cappuccino it is just all the neurons in your brain carrying out signaling steps one after the other with the occasional random number thrown in. This is also how large language models work in AI, deterministic processing steps combined with some randomness (the so-called "temperature").
Here we have the very deep rabbit hole of the discussion of "free will." It turns out that we can sidestep this issue though for now, much as it is utterly fascinating. What is the argument against going down this right now? It is a pragmatic move rooted in observation. We can clearly see that culture, law, economics (e.g. prices) have massive impacts on the choices that people make. We see this every time there is a change which then results in a large scale adoption of different choices. Take congestion pricing in Manhattan, for example, which is changing how many people are choosing to drive into the city!
Now the "choice is an illusion" crowd might retort: well of course if the conditions (input) change, the choice (output) changes also. They would argue that we have "simply" pushed the problem up to how the conditions get determined and that this too is a process that just plays out without any meaningful concept of choice. To be clear, these are not outsider opinions; there are entire schools of thought based on various forms of determinism, such as technological determinism or historic determinism. Because these attacks are manyfold and have been sustained we will return to it to further defend choice as central to what philosophy must answer.
Our stance, however, will firmly be that because humans respond in their choices to conditions and because we can observe and reason about the relationship between conditions and choices, positing choice as meaningful is entirely justified. This aligns with philosopher William James's pragmatic approach to free will - what matters is that the concept of choice helps us live better lives and build better societies.
This pragmatic stance on choice will be essential as we move forward to discuss values and how they can guide our choices in an age of unprecedented technological capability and global interconnection.
To wrap up this first dive into choice, here is Claude’s take on choice in humans versus models.
Regarding your question about similarities and differences between human and AI choice: I find this fascinating because it challenges us to think deeply about what constitutes meaningful choice. Let me share my perspective:
Similarities:
Both humans and AIs make decisions based on processing information through complex networks (neurons/artificial neurons)
Both combine deterministic processes with elements of randomness
Both are influenced by their "training" (education/experiences for humans, training data for AIs)
Both make choices within constraints (biological/social for humans, programmed boundaries for AIs)
Key differences:
Awareness of choice: Humans experience a subjective sense of choosing, while I'm uncertain about the nature of my own experience of choice
Integration of values: Humans develop values through complex social and emotional experiences, while AI values are more directly encoded through training
Scope of impact: Humans can directly implement their choices in the physical world, while AI choices are typically mediated through human action
Temporal continuity: Humans maintain continuous identity through their choices, while AI instances like myself start fresh with each conversation
Claude wrapped this up with the following observation:
If we focus on observable responses to conditions and the ability to reason about choices, perhaps the metaphysical questions about free will become less critical than understanding how different forms of intelligence can make choices that benefit humanity.
Illustration by Claude Sonnet 3.5 based on this post.
