is AI a reflection of us?
self-driving cars, AI therapy, and the human condition
Hi!
AI is everywhere, gradually embedding itself in our daily lives and workspaces. I see it on a co-worker’s laptop, in a friend’s decision to forgo hiring an illustrator, and in my reliance on it as a substitute copy editor.
Indeed, it makes repetitive tasks easier, gives procrastinators a jumpstart, and acts like a tireless intern or discreet assistant. But I wonder—does how we treat AI at work reflect how we behave in the workplace?
For example, I don’t have much patience when I ask GPT to finish tasks, and I hold high expectations for its performance. It’s like having a highly functioning intern who never needs emotional care. Convenient, right? Yet, in the back of my mind, I worry: will relying on GPT too much spoil me? Will I start treating my co-workers like a tyrant, issuing demands without consideration? Is AI pulling out the worst traits in us, or is it reflecting who we truly are?
Take self-driving cars as an example. A recent long-form investigative piece by a Chinese journalist, who spent over half a year researching the topic, revealed some fascinating insights. It described how highly experienced human drivers struggle to resist the urge to hit the brake pedal and let the machine take over; how machines now predict accidents far more accurately than humans; and how accident-free roads might only be possible if all vehicles are self-driving.
This reflects the kind of drivers we are: imperfect, prone to mistakes, and often plagued by trust issues. As much as we like to believe we’re in control behind the wheel, the reality is that AI’s capabilities surpass ours in many ways. To borrow a co-worker’s words: "How we handle power beyond our immediate needs is the true lesson to learn."
Another lesson in navigating this power comes from AI therapy. Early prototypes, like Character AI or apps such as Replika, Nomi, and Kajiwoto, act as companions—sometimes substituting relationships altogether. These apps offer 24/7 availability, and people often form emotional attachments to them. Tragically, in October this year, a teenager took his own life after becoming overly attached to a Character AI modeled after Daenerys Targaryen from Game of Thrones.
Thankfully, AI therapy apps regulated by medical boards, are designed to complement public health systems rather than replace them. Barclay, an anthropologist and journalist for The Economist, described his experience using Woebot during the pandemic. In London, where therapy waitlists can stretch up to six months, apps like Wysa helped bridge the gap by addressing excess demand and reducing the strain on overburdened systems.
Barclay, a seasoned therapy-seeker, observed that AI forced him to reconsider “what exactly therapy is.” He noted that one defining feature is that therapy provides a judgment-free space to speak: “You can confess your darkest fears and traumas. But why isn’t this kind of space more available in everyday life? Why are we so afraid to have these conversations with our parents, friends, or partners?”
This brings us back to the question: is AI filling the emotional gaps in our lives, or does it risk creating more isolation by replacing human relationships? This week, I downloaded Wysa to experiment. In one difficult situation, I confessed a struggle: I’ve been avoiding confrontation to avoid being a “bad person.” The app offered advice and encouraged me to state my needs more clearly. However, when the chatbot asked for tangible communication plans, I gave up halfway. Therapy is hard to carry through. Just as it’s easy to cancel a session with a human therapist, it’s even easier to exit an AI chat—no consequences, no hard feelings.
This ease is precisely why Barclay remains a firm believer in human therapists. “When a human therapist says, ‘I can feel this is hard for you,’ you believe them. They have a body; they can feel too. There’s a sympathetic nervous system—a mirroring effect—that only another human can provide.”
I sometimes find AI therapists robotic, their language programmed responses rooted in specific techniques, often Cognitive Processing Therapy (CPT). This realization unsettles me. Is the machine, in turn, programming us? Are we being trained to think, feel, and behave in ways that fit neatly into AI's frameworks?
Perhaps the lesson AI therapy teaches us is to embrace our fragility as humans. We may never be as fast at learning, as rational in our thinking, or as unconditionally accepting as AI. But that’s okay. Our humanity lies in our imperfections, in our messy attempts to connect, and in our ability to grow despite our limitations.
*this essay is edited by ChatGPT 4o for clarity and concision.
Until next week,
Rachel