Jump to section
The Big Shift
When people think of AI in mental health, they imagine — or worry about — chatbots replacing therapists. That's not what's actually happening.
In a recent Bloomberg Television interview, April Koh, CEO and Cofounder of Spring Health, pointed to something more interesting: therapy and emotional support have become AI's fastest-growing use case. The number more than doubled in 2026, jumping from 5% of all AI applications to 11%. People aren't replacing human care. They're demanding mental health support in a different way.
"People want to consume mental health care in a different way," Koh explained to Romaine Bostick and Katie Greifeld, hosts of Bloomberg Television’s “The Close.” "And AI is bringing on a real mental health revolution, and changing the way people are experiencing mental health care."
But here's where it gets complicated. General-purpose AI tools aren't built for this. They're optimized for engagement and productivity, not for the clinical realities of supporting someone in distress. That's why Spring Health built its AI differently.
Built for safety, not just scale
When Spring Health looked for industry standards to guide safe AI deployment in mental health, they found nothing. No benchmarks. No frameworks. No agreement on how to evaluate whether an AI system could actually handle suicide risk detection or know when to escalate to a human clinician.
So they built one. VERA-MH, the first open-source benchmark for evaluating ethical and responsible therapy delivered through AI, emerged from collaboration with the AI in Mental Health Safety and Ethics Council, a cross-disciplinary council of academic, healthcare, and technology leaders. It's become Spring Health's north star.
"We not only hold ourselves accountable to these standards," Koh said, "but we've open-sourced this benchmark for the industry. It's not proprietary. It's a gift."
The difference matters. Spring Health's AI lives inside HIPAA and SOC 2 compliance. It's anchored to clinical networks and safety protocols that general-purpose LLMs simply don't have. Member data never trains external models. The guardrails are built in, not bolted on.
Expanding access without cutting corners
The interview also touched on Spring Health's new neurodiversity program: specialized care for children with neurodivergent needs that's been embraced by employers facing a real gap in their benefits.
"Our mission is to eliminate every barrier to mental health," Koh said. "Family is a huge part of that. We're in an acute mental health crisis that's only growing, and neurodiversity is a massive part of the demand."
The program reflects a core principle: scale doesn't have to mean less personalization. It means better matching: using AI to connect each person to the exact right support from the start.
Built to last
Spring Health is profitable and self-sustaining — a rare position for a company in behavioral health at this stage. That independence shapes how the company thinks about growth.
"We're building for the long term," Koh said when asked about the company's public market timing. "We will go public when the timing is right. Right now, we're profitable, so we're not capital constrained. We're very opportunistic about how we deploy capital, but on our own terms."
It's the kind of position that lets a company say no to the wrong deals and yes to the ones that matter.
About the interview
This Bloomberg Television segment is part of a series examining how artificial intelligence is reshaping healthcare sectors. The interview was conducted with April Koh as part of Bloomberg's ongoing coverage of AI's expanding role in mental health and behavioral health innovation.
Spring Health, a global mental health company built on the first AI-native mental health platform at scale, supports more than 170 million lives through employers and health plans. The company was founded in 2016 as an AI company by April Koh and Adam Chekroud, informed by machine learning research in psychiatry.
.avif)
