Why Human-Centered AI Is the Future of Medicine & Mental Wellness

The expert ties this to the concept that social skills, like a muscle, weaken if not used. She models the protocol herself and insists her children follow it. Because voice assistants and chatbots are becoming conversational partners, she believes that how we engage with them reflects and reinforces our social norms. She also tells her kids that rudeness to Alexa can translate into rudeness to people, so the practice is as much about preserving humanity as it is about etiquette.

El Kaliouby's 16-year-old son is actively trying various AI tools. She uses his curiosity to spark discussions about the system's limitations: why it answered the way it did, what training data shaped its response, and how to question its output. She applies this philosophy in her school board role, advising against blanket bans on AI. Instead, she advocates for intentional integration that teaches discernment. The approach prepares children for a future where AI is ubiquitous, ensuring they can leverage it without being misled by misinformation or algorithmic bias.

The expert describes a future 'health co-pilot' that aggregates data from the Oura ring, continuous glucose monitors, and even a hydration patch that measures sweat electrolytes. With natural language processing, a user could ask, 'Given my sleep and nutrition over six months, what should I change?' Similarly, a clinician could query a patient’s multi-omics data as if talking to a PhD student. El Kaliouby sees the technological pieces as ready, but no one has assembled them into a seamless experience. She also envisions agentic AI that could autonomously book doctor appointments or manage supplement schedules. This protocol reflects emerging possibilities rather than a currently available tool.

El Kaliouby served on the board of Videa Health, which uses natural language processing and emotion AI to analyze video interview responses. The app flags detachment, depression, or anxiety and has reportedly helped prevent suicides. She connects this to broader societal issues like gun violence, suggesting that if technology could detect dangerous emotional trajectories early, it could divert individuals from harming themselves or others. She acknowledges that the data-sharing model (who gets alerted) is unresolved, but the technology is ready.

El Kaliouby describes being newly married and in a long-distance relationship, sobbing in front of her laptop while chatting with her husband. She could only send a crude emoji, which sparked the question: what if technology could read and respond to non-verbal signals? At the time, computer vision wasn't applied to human emotion. She began building algorithms to map facial expressions to emotional states, effectively creating 'emotion AI'. This was not only technically challenging but also culturally provocative—many tech people initially dismissed the idea that machines should understand emotions. Yet she points out that emotional intelligence (EQ) is what makes leaders effective, and the same principle should apply to AI. This personal origin story underscores why she believes human-centric AI is so vital.

The expert argues that constant mediated communication, especially among children who grow up with smartphones, weakens the ability to read nuanced emotional expressions or make eye contact. She compares this to a muscle that atrophies when not used. The result is a society where people are more polarized and detached, and where even tragic events become opportunities for filming rather than intervening. However, she doesn't advocate for abandoning technology; instead, she believes AI must be designed to rebuild emotional connection. This dual perspective—acknowledging tech's role in the crisis while insisting it can be part of the solution—is her central thesis. She insists that AI companions, bots, and future systems must have emotional intelligence to counter dehumanization.

The expert describes how MIT’s new AHA program (Advancing Humans with AI) and Stanford’s HAII (Human-Centered AI Institute) are institutionalizing this approach. She imagines defining four or five measurable KPIs around 'human flourishing'—such as better relationships, increased empathy, or improved mental health—and using them to guide AI development. She admits that human flourishing is nebulous and hard to quantify, but she insists it’s a necessary direction to avoid a dystopian future. She also practices this as an investor: when evaluating startups, she checks for commitment to ethical development, like data diversity and consent protocols, treating ethical consideration as a green flag for investment.

El Kaliouby envisions a health co-pilot that combines data from wearables (Oura ring, CGM, hydration patches) and interprets it through predictive models, while a generative interface allows natural language querying—e.g., 'Given my sleep data over six months, what should I change?' Clinicians would similarly interrogate patient data as if talking to a PhD intern. She believes the pieces exist but lack a unifying platform, an 'iPhone moment'. This integration could finally solve functional medicine's data overload problem. She predicts this will happen within a couple of years, followed by agentic AI that books appointments and manages supplement schedules autonomously.

She argues that who writes the check determines who gets to build the future. Without more women and diverse investors, AI will keep solving the same problems defined by a homogeneous group. Her personal history—from Egypt, a post-doc at MIT, a successful exit—shows that minority perspectives can create entirely new markets (emotion AI). She now actively seeks out founders building in women’s health, mental health, and healthspan, areas historically underfunded. She also mentors her own children to be critical consumers of AI, emphasizing that the responsibility of diversity extends to education at home.