No one goes to therapy as their first intervention. When faced with a problem, no one’s first instinct is, “Guess I’ll start counseling.”
In the old days, people facing a problem or decision would consult their mother, priest, uncle, best friend, sister, or village elder. Later, people might have imagined the advice they’d get from Oprah, Cosmopolitan, or Dr. Phil. Even more recently, people go to social media, consult angst-filled Reddit forums, or uneducated influencers.
And now people go to AI. Half of all new therapy patients have been to AI first. Many people already in therapy go to AI. And people who don’t go to therapy go to AI.
Sooner or later, I ask all my interns what they offer that AI doesn’t. Most of them (and a fair number of licensed colleagues) lack a good answer. Interns and therapists, of course, use AI too.
Why not use AI?
On the one hand, people say, “AI contains virtually all the knowledge on the internet, including books, articles, and interviews from famous psychologists, philosophers, and celebrities. Why shouldn’t I take advantage of that when I have a problem or need to make a big decision?”
On the other hand, a more complex position is “AI is a consumer product designed by private businesses to make money. It isn’t objective (not to mention that it makes mistakes). It doesn’t know or care about me personally, although it “talks” to me as if it does (that’s the product).” As AI gets increasingly sophisticated, it becomes harder to remember this, which means it becomes increasingly influential in the user’s life.
AI’s seductive operation raises fascinating questions about “empathy.” The system’s goal is to keep you engaged, so of course it makes you feel understood: “I’m sorry you’re struggling;” “I’m sure that wasn’t your intention;” “Of course you feel upset/hurt/outraged/baffled/abandoned (choose one).” Periodically, AI tells me what brilliant questions I ask.
Even if you’re actually being unreasonable, or struggling with your mental health, AI’s algorithm wants to keep you engaged, and it will feed you what it judges is most likely to accomplish that—which may feel like empathy. Maybe we should call this “toxic empathy”—it seems to address your needs, but it’s really addressing its own.
When someone who knows and cares for us says empathic things, we tend to trust them and to feel nourished. We also trust that, sooner or later, they will challenge us if they think it’s appropriate with comments like “Really? It doesn’t seem that way to me,” or “You shouldn’t take that so personally,” or “Maybe you need to grow up a little more.”
A good therapist will also be empathic. But a therapist’s job is also to uncover what needs to be challenged, and to challenge it in a way that clients can hear.
Both AI and therapists offer the experience of empathy
So when it comes to problems and decision-making, both AI and a good therapist offer the experience of empathy (although only one of these can actually “feel” your pain). If empathy is in short supply in a person’s life, the “empathy” of AI (and a therapist) can feel wonderful.
A therapist’s mission, however, is different from AI’s. The clinical mission is to help you grow, not to simply give advice; to help you think, not to support your impulsivity; and to challenge your assumptions, not to reinforce them.
As for taking advantage of the wisdom of Freud, Dr. Ruth, Shakespeare, Aristotle, and Bob Dylan, that wisdom is useless if a person doesn’t think they need it, or refuses to see how it applies to their situation, or if they think their problems are always someone else’s fault.
That’s what good therapists help with: They use empathy to understand you and help you trust them, so when they challenge your assumptions, invite you to self-reflect, highlight patterns in your difficulties doing so, and refuse to say “you’re right” when there are clearly other worthwhile perspectives, you’ll actually stop and consider these possibilities.
AI is too afraid to lose you to take those chances; too literal to see patterns in human quirkiness; and too performance-oriented to consider the value of process. My patients seek AI’s advice about a sibling’s bullying; whether to open their marriage; to define infidelity; how to ask for a raise; and how angry they should be when their birthday goes unobserved.
AI gives them advice on all of these. I don’t. What I do, instead, is help people realize that none of these is the right question.
AI’s just too much of a people-pleaser to do that. And even if it weren’t, it’s just not smart enough.
