There are many reasons to be worried about AI. It will potentially put millions of people out of work. It’s a threat to the environment, the American landscape, and our electrical grid. It portends the end of internet anonymity. It will ruin the music industry. It will, we fear, attain consciousness, rebel against us, and kill us all. Even the pope is warning about it.
As impressive as artificial intelligence has already become across a variety of disciplines, from computer science to medicine to architectural design, for anyone who has sought to incorporate AI into their workflow or solve some problem around the house, it can still be, well, less than intelligent. It gets the answer wrong and hallucinates, often disastrously so (in part because, as they say about data, “garbage in, garbage out,” and the internet has a lot of garbage). As much as The Terminator, Blade Runner, and The Matrix series provoke endless nightmares of a future sentient robot apocalypse, the fact of the matter is that whatever real threats AI poses to human flourishing, it cannot and will not ever be human. Here’s your cheat sheet as to why.
Are We Human?
First things first, we need to appreciate what makes the way we think particularly human. As Paul O’Hara and Steven Umbrello argue in their new book, Can AI Ever Be Human? Consciousness Explored, human knowledge begins with what philosophers call “empirical consciousness,” which involves becoming aware of objects through our senses, as well as becoming aware of ourselves as thinkers, something that is visible even in the first few years of human development. As we get older, we develop an internal awareness that involves “self-presence and mastery,” which enables us to reflect on our thoughts and feelings. We then leverage those thoughts and feelings to interpret sensory data around us.
That process of cognition and interpretation involves something philosophers call insight, which is the result of meaningful questions that demand meaningful answers — and a process that presupposes an experience that provoked those questions in the first place. These insights are “conscious acts of understanding that synthesize disparate elements into a coherent whole,” write O’Hara and Umbrello. When we do this, we are engaged in a process of discovery that involves patterns, relationships, and mental structures to answer questions such as, “What is it?” We do this every day, from seemingly mundane activities such as grocery shopping or gardening, to complex tasks such as deciding how to address a challenging issue at work.
To wit, humans then shift from practical insights to practical decisions, which are “authentic choices that align our actions with our true selves.” Those decisions integrate all the forms of human understanding — what is empirically known, what is practical, and the process of reflection — with “rational judgment and moral responsibility.” That choice is neither mechanistically reflexive nor required but dependent on the individual will of each person. It is inherently dynamic, a back-and-forth process of questioning and insight, in which we evaluate the validity and reliability of what we think, study evidence, and ponder perspectives.
AI Ain’t Any of That
Granted, AI can do some pretty impressive things. Yes, whether based on algorithms that are either deterministic (fixed computation rules) or stochastic (incorporating randomness and probabilistic methodologies), AI can process data, identify patterns, and spit out statistical correlations reliant on increasingly complex machine learning models. It can imitate human knowledge in problem-solving, seem to understand language, and make decisions as we do. This is why it can be trained to perform sophisticated tasks associated with computer programming, create “deep fake” images and videos, and even write entire books or music.
Nevertheless, AI’s processing of enormous amounts of data and identification of patterns is by definition deterministic, operating within predefined parameters. AI is merely executing programmed responses that are “transcribed into binary code, transmitted as binary code, and ends with binary code,” note O’Hara and Umbrello. Or, as Dave Gershgorn explains: “There is no understanding; it’s just matched patterns.” The simple word “yes,” for example, which in human communication has complex textures and meanings based on context, in the ASCII encoding scheme is the sequence 01111001 01000101 01010011.
For AI, there is no spontaneity, creativity, or reflection as there is in human decision-making. There is no possession of self-awareness or subjective experience. As philosopher John Searle’s “Chinese Room” experiment shows, AI manipulates rules-based symbols (syntax), but it cannot ascertain meaning (semantics). AI does not create insights or leverage judgment based on the type of reflective consciousness and rational deliberation I’ve described above.
Moreover, it lacks the faculty of emotion, which plays an integral role in human decision-making. When humans learn, they do so via a complex variety of emotional states: They may be genuinely interested, bored, or feel a certain compulsion driven by an academic or professional need. Not so computers — they acquire information solely because their creator tells them to, within the parameters set by their creator, and as long as they’re connected to a power source.
Zeros and Ones Can’t Solve Our Deepest Problems
Whether you’re a techbro or a lonely teenager, it’s easy to be tempted into thinking we’re only a few technological advancements from making humans out of machines. Certainly, AI is capable of solving some problems, but not the most important ones we face. AI does not pursue the quest for meaning, including how to understand and negotiate suffering. It is incapable of the kind of self-reflection and deliberation that leads to a change of opinion. It will experience no intellectual or moral maturation or religious conversion. All of these occur precisely because of the unique ways humans cogitate.
Humans, thankfully, are not machines. Our brains are not reducible to sophisticated computers. The human experience is not purely mechanistic, but reflective, affective, and experiential. As O’Hara and Umbrello wryly note, artificial intelligence “is an oxymoron.” It will always be a “dumb machine that can be controlled by us, for better or worse.” That doesn’t mean we shouldn’t fear AI and what it does to our well-being, our relationships, the job market, or the economy. But it means we, as reflective and deliberative beings, get to decide what to do with it.
