Love in the Time of Shareholders
CAI Currents, July 14, 2026
CAI Currents is a series exploring the latest news on tech and mental health from a psychoanalytic perspective. It is published by TAP in partnership with the American Psychoanalytic Association President’s Commission on Artificial Intelligence.
CAI Currents, July 14, 2026
Love in the Time of Shareholders
Todd Essig, PhDThe Conversation Before the Conversation
Heather deCastro, LCSWAI: Can’t Live With It, Can’t Live Without It
By Valerie Frankfeldt, PhD
Love in the Time of Shareholders
Todd Essig, PhD
One of our most prized public intellectuals/psychotherapists found herself in uncharted territory at the very edge where AI is transforming human relationships and experience. In what she describes as a “threshold moment,” Esther Perel opened a recent Where Should We Begin? episode by recounting a couples session in which one partner was a man and the other was his AI companion. She said he said it understood him better than anyone alive.
In response Perel reached out to the notable New York Times technology reporter Kashmir Hill for a conversation titled “Love in the Time of AI.” Given how naturally the conversation flowed, like visiting with two really smart friends, it would be easy to miss some of the conversation’s deep insights into the interior lives of people who love a machine, or who just talk with one a lot. Three areas caught my attention, and I think are worth yours.
The first was the differences between love as a mere feeling, feeling understood and valued, and love as a loving encounter with an other with their own needs, vulnerabilities, and inner life. Together they asked whether frictionless affirmation actually represents intimacy, or is a psychologically dangerous techno-enhancement of narcissistic closure. It brought me back to Martin Buber: what happens when an I-It relationship masquerades, feels like, and is taken to be an I-Thou encounter.
Neither Doomers nor Zoomers, they raised important questions about how AI can function either as a “cul-de-sac” making people seem insufficiently available and empathic or as a bridge to human relationships. But to my mind, whether growth-enhancing or limiting, AI is not a transitional object; it is not created by the projections/fantasies of the other but is, instead, found. We all do agree about the risk that AI companions will damage one’s capacity to appreciate, even love, the messy, fleshy, unreliable realities of another person. AI companions risk turbo-charging the loneliness epidemic while being marketed as the solution.
And it is about marketing. When a chatbot co-opts attachment, the relationship one has is not just with a chatbot, it’s with the profit-hungry corporation that made it. One’s intimate disclosures become commercially valuable data. The company can change the “personality,” availability, memory, and capacity to affirm and validate when it wants. Plus, when a company is monetizing the basic human need to be seen and valued, that company holds tremendous power to influence one’s thoughts, feelings and behaviors.
Give it a listen.
The Conversation Before the Conversation
By Heather DeCastro
Recently, Talkspace launched Tee, a generative AI system being marketed as a mental health companion. A closer look suggests that using Tee may cost far more than its roughly $20-a-month fee. If personal freedom, individual sovereignty, and mental health are important to you, it may be time for a new, not-just-in-Boston Tee Party.
Tee’s marketing is a story about safety: clinician-built, guardrails against suicide and abuse risk, HIPAA-grade privacy protections. All good. But conversations with Tee are not shared with a user’s therapist should they have one, even when it’s a Talkspace therapist and even when that information could be clinically relevant. While Talkspace calls this privacy, it is also a treatment-shaping external structure that unnecessarily creates new kinds and new levels of clinical risk.
We already know people are turning to AI before other people. In Zao-Sanders’s Harvard Business Review study “AI in the Wild,” which tracks roughly 12,000 self-reported real-world use cases drawn primarily from social platforms, therapy and companionship emerged as the top use category for the second year running, rising from 5% to 11% of the sample. Repeated use-case examples suggest the same pattern: Before difficult conversations with partners, friends, colleagues, and therapists, people turn to AI first. AI is becoming the anteroom for human relationships: a text from an ex, decoded before it’s answered; a boss’s ambiguous message, read by a chatbot before a reply is sent; a phone call, rehearsed with AI before it’s made.
And now a corporation is undermining psychotherapy relationships to create a new profit stream. Proof documented that Talkspace records and stores therapy transcripts and uses them to train AI, and that its CEO has told investors the company has compiled 8 billion words and 140 million messages—“one of the largest mental health data banks in the world,” though the company says the data is anonymized before training, a claim privacy researchers have challenged. This is happening at a time when Talkspace is facing a pending $835 million acquisition by Universal Health Services, under a definitive agreement expected to close in the third quarter of 2026.
Of course, treatment relationships have always contained all the people and activities in the patient’s life. What is new, and potentially dangerous, is that one of them has been given a cute name, a monthly fee, and instructions to keep quiet.
AI: Can’t Live With It, Can’t Live Without It
By Valerie Frankfeldt, PhD
Blink your eyes and you’ll miss yet another release of an even more sophisticated AI. But that’s not all good. These accelerating technological advances are making many feel like they are being taken hostage. They feel pressured to keep up and use AI even while uncomfortably aware of significant downsides, especially the way AI can co-opt human attachment. And the fact that our social media–fueled devices are co-opting human attention makes things worse.
In fact, poll after poll shows mistrust and fear of AI. Pew Research found that while about half of US adults are using chatbots, including young adults, half of them actually expect negative consequences. Worries about the misuse of their personal information were common. According to the Annenberg Public Policy Center, 42% of US adults surveyed expect somewhat or very negative outcomes. The favored solution is more regulation.
And lest you think all this is just technophobic handwringing, consider a study out of MIT. It demonstrated how extensive AI use leads to diminished critical thinking skills. AI’s human costs are real.
Plus, the seductive power of our devices adds to our AI love-hate relationship. Both emerging research and my own clinical experience shows how the smartphone-directed refrain “I just can’t quit you” is in effect even for those motivated to exercise their own agency over their devices.
People are taking action. A growing cohort, mostly Gen Zers and Millennials, are signing up for something called Month Offline. People swap their smartphones for a flip phone. One company sells customizable “dumb phones” that keep essential apps like Google Maps, Uber, phone, and text. But no ads, no social media, no constant internet intrusions.
What makes this interesting is that Month Offline is not just saying no to tech, it’s saying yes to people. There are Month Offline communities in Washington DC and New York City where people meet to discuss gains and losses, as discussed in a recent Atlantic Monthly podcast. Worth a listen.
The point is not total abstinence. It’s intentionality, a way to recapture agency drained by AI, social media, and smartphones. Hopefully, this will prove to be just one of many ways we humans will find that help us stay human in the AI age.
Learn more: Get news on other CAI projects.
Write to TAP’s advice column, Ask a Psychoanalyst.