Despite widespread mythmaking about LLMs (generative AI tools like ChatGPT, Claude, and Gemini) replacing workers, this fear is overblown. While these tools boost productivity by automating mundane tasks—causing some information workers to hide their usage from bosses—the reality is that current workplace AI usage data represents only a small segment of the labor force. The more compelling question isn't about job replacement, but rather: how are people using generative AI in their personal lives?
New evidence suggests people use generative AI not just for productivity enhancement, but to address fundamental human needs. In the April 2025 issue of the Harvard Business Review, a new report tells us that in 2025, people now turn to LLMs for therapy (#1 overall), life organization (#2), and even finding purpose (#3). You read that right: therapy and companionship emerge as the primary use of generative AI this year. Far from boosting productivity, generative AI unexpectedly serves a role as companion and confidante.
I admit reading the report and feeling starry-eyed. I've argued that generative AI marshals us into a new revolutionary age that will allow us to live longer, better, and more meaningful lives. From helping us achieve our most audacious "moonshot" ideas, to transforming us into Nobel-worthy innovators, and now to providing much-needed mental health support, generative AI proves there's a path to bettering ourselves and improving the human condition.
I think this reveals a striking truth about our technological moment: we're turning to machines not just for calculations, but for hope.
Does this represent a true advance? Or a gilded cage? While generative AI offers benefits like 24/7 availability and low barriers to expertise, lurking beneath the surface are growing concerns about data privacy, political bias, and the potential for over-reliance. "I cancelled my subscription for this reason. I do not want to support a tool that tries to turn the world more timid and prim,” states one person in the report, illustrating a budding skepticism. I've personally heard an even more jaundiced view: why seek help from machines that just regurgitate the internet?
Here we see a familiar pattern: overt techno-optimism followed by a healthy dose of reality.
It's fair to express some skepticism. But no amount of skepticism changes the reality: outside of work, people who use generative AI long for companionship and expertise. It shouldn't surprise us: the traditional mental health services model remains catastrophically inaccessible for most people worldwide. As one South African interviewed for the report puts it: "There's a psychologist for 1 in every 100,000 people and a psychiatrist for 1 in every 300,000 people." With fewer professionals and rising demand, how can this traditional model survive?
Here's another reality: simple economics drives this hope. Generative AI is available round-the-clock, affordable (often free), and meets the most basic, profound human needs, all with just an internet connection. The same South African tells us "large language models are accessible to everyone, and they can help." In other words, where traditional systems fail, AI picks up the slack.
To be sure, mental health expertise is but one of an array of professional services out of reach to most humans (think tax advisory, strategic consulting, legal services, to name a few). Against a backdrop of scarcity, generative AI offers a hopeful alternative to a model that's simply unaffordable to most people. This is where the mythmaking behind LLMs start to shape reality.
Is this a transformative improvement? Absolutely. We already see generative AI making our work lives more productive; now we see evidence that it brings companionship and expertise into our personal lives. But does it actually bring hope? At the risk of sounding Pollyannish, the techno-optimist in me thinks it does.
Tony McGovern
Tony McGovern is Founder and Data Scientist at emdata.ai.
