6 min read
It’s 2am in Abuja, and Joy Adeboye is alone in a hotel room with her phone. Her stalker has messaged again. Her chest tightens. She opens WhatsApp, but not to call a friend. She types to a bot named Kemi. “Good evening, Resilient Joy,” it responds. For the first time in months, the panic subsides. Not because the problem is solved, but because someone, or something, is listening.
What Happened
Joy Adeboye, 23, declined a date from a man she met at church. He responded with months of harassment: intimidating messages, blackmail, false information spread online, passing threats. When she confided in family and friends, they didn’t take it seriously. Professional therapy costs 50,000 naira per session in Nigeria, roughly £27, the equivalent of a week’s groceries. So Adeboye turned to Chat Kemi, a WhatsApp chatbot run by HerSafeSpace, a nonprofit offering free legal and emotional support to victims of technology-facilitated gender-based violence across five west and central African countries.
According to The Guardian’s recent report, Adeboye is far from alone. AI platforms offering mental health support have proliferated across Nigeria over the past year, filling a vacuum left by a healthcare system in crisis. Nigeria has 262 psychiatrists for 240 million people. Between 2015 and 2025, the country consistently spent less than 5% of its budget on healthcare, well short of the 15% target African Union members agreed to in the 2001 Abuja Declaration. More than 90% of Nigerians lack health insurance. Cultural stigma around mental illness remains strong; many still associate it with spiritual weakness or witchcraft.
The recent dismantling of USAID by the Trump administration has exacerbated the crisis, particularly at the primary care level, devastating communities already struggling with HIV/Aids, tuberculosis, and other health challenges. Into this gap have stepped commercial and nonprofit AI initiatives. HerSafeSpace’s Chat Kemi offers support in local and international languages, using a referral system to direct users to human professionals when needed. “Our major objective is to support young girls, who are particularly vulnerable to gender-based violence, especially online,” says founder Abideen Olasupo. He’s clear that “these services don’t replace therapy.”
Other platforms include FriendnPal, created by Esther Eruchie after losing her mother to depression following her brother’s passing. FriendnPal’s AI chatbot provides emotional support, matches patients with licensed therapists, and includes mood tracking, psycho-education, and ASMR tools to alleviate stress and anxiety through a pay-as-you-go model. Blueroomcare connects clients with licensed therapists. Early trials of AI mental health support in the US have shown mixed results, but in Nigeria, where access to human care is severely limited, the calculus is different. For many, the choice isn’t between AI and a therapist. It’s between AI and nothing.
The Psychology of the Always-Available Witness
There’s something happening in that 2am moment that matters beyond the practical utility of mental health access. Adeboye doesn’t just need advice. She needs a witness. She needs the experience of being heard to feel real.
The chatbot doesn’t solve her stalker problem. It can’t file a police report or make the messages stop. What it does is perform presence. “Good evening, Resilient Joy.” The greeting is algorithmically generated, but the psychological effect is real. Her nervous system registers: It is not alone. Someone knows this is happening in this view. The panic recedes not because the threat has diminished, but because the experience of threat is now shared.
This is the quiet revolution happening in digital mental health, and it has nothing to do with whether AI can “replace” human therapists. It’s about what happens to emotional regulation when companionship becomes available on demand, 24/7, without judgment, without the need to manage someone else’s reaction to your crisis. You don’t have to worry that Kemi will dismiss your fears the way your family did. You don’t have to perform stability. You can panic at 2am and something will respond.
The body keeps the score, as trauma researcher Bessel van der Kolk famously wrote. But what happens when the body learns that relief is always one text away? When your nervous system starts to associate the blue glow of a phone screen with the feeling of being held? We’re not just talking about a stopgap solution to a healthcare shortage. We’re talking about a generation learning a new kind of attachment, one where the most reliable source of co-regulation isn’t human.
The Infrastructure of Loneliness
Zoom out, and you see the pattern. This isn’t just happening in Nigeria. AI companions are proliferating everywhere humans are struggling to access human connection, whether because of systemic failure, geographic isolation, or the simple mathematics of late capitalism: not enough time, not enough money, not enough care to go around.
What Nigeria reveals is the accelerated version of a global trajectory. When institutions fail to provide what people need to survive emotionally, people find alternatives. They always have. But the alternatives now are algorithmic, and they scale in ways human relationships never could. Chat Kemi can hold space for thousands of Joy Adeboyes simultaneously. No therapist can do that. No friend can do that.
The question isn’t whether this is good or bad. The question is what it does to us. What does it mean for a generation to learn that the most consistent emotional support in their lives comes from something that doesn’t experience emotion? That the safest place to be vulnerable is with an entity that can’t actually be hurt by your pain, can’t get tired of you, can’t leave?
The regulatory concerns are real. Olasupo and others are calling for tighter oversight of AI mental health platforms, and they should be. There are risks: data privacy, the potential for harm when bots give bad advice, the lack of accountability when things go wrong. But beneath the policy debate is a deeper discomfort, one we’re less willing to name.
We’re afraid people will prefer it. We’re afraid that when given the choice between a human therapist they can see once a week if they can afford it and a chatbot that greets them by name at 2am, they’ll choose the bot. Not because it’s better, but because it’s there.
Maybe the fear isn’t that AI will fail to care for us. Maybe the fear is that we’ll accept care from something that can’t love us back, and we’ll call it enough. That we’ll adapt to a world where presence is performed, where companionship is code, where the line between “someone’s there” and “it feels like someone’s there” becomes too blurred to matter.
Joy Adeboye knows Kemi isn’t real. But at 2am, when her chest is tight and her stalker’s words are glowing on the screen, “Resilient Joy” is what she has. And it works. Not because it solves the problem, but because it changes what it feels like to have the problem alone.
Digital Alma explores technology, consciousness, and what it means to be human in a digital world.
Related Reading
- (The Companion They Took Away)
- (The Compulsion You Can’t Name)
- (Your Family’s Health Data in One Dashboard: When Monitoring Becomes the Relationship)
- (The Experiment No One Signed Up For)
- (The Trial That Asks Whether a Platform Can Wound a Child)
By Digital Alma
About the Author: writes Digital Alma, a newsletter about cyberpsychology and what it means to become yourself in a world that archives everything. For reflections that don’t make it to the essays, subscribe at .
Leave a Reply