How AI Can Help Fix Our Brains

TL;DR: Today we’re releasing a new episode of our podcast AI & I. Dan Shipper goes in depth with Awais Aftab, psychiatrist, professor, and writer. Watch on X or YouTube, or listen on Spotify or Apple Podcasts. Here’s a link to the episode transcript.

Was this newsletter forwarded to you? Sign up to get it in your inbox.

The same rigid thinking that held AI back for years is failing millions with mental health conditions—but AI’s progress offers a radical solution.

For years, computer scientists tried to teach machines using fixed rules, but that approach crumbled against the messy reality of the real world. Psychiatry faces the same trap today, forcing the infinite variations of human suffering into neat diagnostic boxes—reducing someone's unique experience of depression or anxiety to items on a checklist.

AI's breakthrough came from embracing deep learning—letting computers discover their own patterns from examples, rather than forcing predetermined rules. Now that same approach could transform mental health care: AI systems that recognize how symptoms cluster differently in each person and catch disorders before they fully develop.

In this episode of AI & I, Dan Shipper—whose interest in this topic arose from his own experience living with obsessive-compulsive disorder (OCD)—explores this transformation with Awais Aftab, who has been questioning psychiatry’s rigid categories from inside the field. Aftab is a clinical assistant professor at Case Western Reserve University, editor of Conversations in Critical Psychiatry—an Oxford University Press volume that tackles philosophical and critical perspectives in psychiatry—and author of the Substack newsletter Psychiatry at the Margins. You can check out their full conversation here:

If you want a quick summary, here are some of the themes they touch on:

How AI could map the landscape of mental disorders

Dan thinks that psychiatry is facing the same problem that early machine learning researchers did. For example, the rules they used to train a computer to recognize the letter “A” might’ve broken down because the letter can vary dramatically in shape, size, and style depending on handwriting or font.

The real breakthrough came with deep learning, where neural networks are fed examples and allowed to form their own patterns of recognition. Dan explains, “No individual neuron knows like, hey, this is like an ‘A,’ but altogether… the network is more or less testing out a bunch of different hypotheses for what the letter could be."

Similarly, psychiatry often tries to categorize disorders like OCD using a diagnostic checklist of sorts. But this approach misses the nuances, as symptoms often blur across categories. “[OCD’s] boundaries overlap with other anxiety disorders, overlap with depression… and biologically speaking, there is no one single thing that makes OCD, OCD—but rather its variety; so we have to respect that heterogeneity,” Aftab says.

Aftab sees potential in using AI to explore new ways of mapping mental health symptoms—what he calls the “space of psychopathology.” Rather than hoping to find a single clear cause or “essence” for disorders, Aftab prefers to view mental health as “a big fluid fuzzy mess of interacting causes,” which can be carved up in many different ways. AI could help us discover hidden patterns or groupings of symptoms, allowing psychiatry to better understand the complex reality of human minds.

Why psychiatry should stop explaining—and start predicting

Taking this idea further, Dan argues that science tries too hard to explain why things happen rather than focusing on predicting what will happen next. He contrasts this to how LLMs operate: We know they predict each word based on what came before, but exactly how millions of parameters interact to produce accurate predictions remains opaque.

Dan suggests psychiatry adopt a similar approach with disorders like OCD. Instead of endlessly debating its underlying causes, he says, “We should just gather tons and tons and tons of both contextual data and biological data and chat logs and just throw it into a deep learning model to predict.” Aftab notes that similar predictive approaches have been attempted in studies of suicide risk, sometimes achieving impressive accuracy—but often breaking down beyond the sample datasets they were trained on. He remains optimistic that gathering significantly larger datasets could eventually overcome this hurdle, mirroring how image and text generation improved dramatically with more data and compute.

How ChatGPT enables better, more accessible mental health care

Aftab also sees an immediate use case for AI in the field of psychiatry: AI-assisted clinical interviews. He notes that clinicians sometimes lack the time and attention to catch nuanced details in their conversations with patients. An AI clinician, however, “would have the resources to go into the nitty-gritty of the symptoms,” clarifying and exploring nuances a busy human practitioner might miss. This could improve patient care by helping clinicians identify overlooked patterns early.

Dan adds that many people already use ChatGPT to do this. “It democratizes access to the most basic level of mental and emotional support,” he says, helping users manage issues before they escalate into larger problems. Dan imagines a future where AI could guide patients with clear suggestions about what to discuss in their therapy sessions, making human-led treatment more effective and targeted.

Here’s a link to the episode transcript.

You can check out the episode on X, Spotify, Apple Podcasts, or YouTube. Links are below:

1. Watch on X
2. Watch on YouTube
3. Listen on Spotify (make sure to follow to help us rank!)
4. Listen on Apple Podcasts

What do you use AI for? Have you found any interesting or surprising use cases? We want to hear from you—and we might even interview you.

Miss an episode? Catch up on Dan’s recent conversations with founding executive editor of Wired Kevin Kelly, star podcaster Dwarkesh Patel, LinkedIn cofounder Reid Hoffman, former a16z Podcast host Steph Smith, economist Tyler Cowen, writer and entrepreneur David Perell, founder and newsletter operator Ben Tossell, and others, and learn how they use AI to think, create, and relate.

If you’re enjoying the podcast, here are a few things I recommend:

1. Subscribe to Every
2. Follow Dan on X
3. Subscribe to Every’s YouTube channel

Rhea Purohit is a contributing writer for Every focused on research-driven storytelling in tech. You can follow her on X at @RheaPurohit1 and on LinkedIn, and Every on X at @every and on LinkedIn.

We build AI tools for readers like you. Automate repeat writing with Spiral. Organize files automatically with Sparkle. Deliver yourself from email with Cora.

We also do AI training, adoption, and innovation for companies. Work with us to bring AI into your organization.

Get paid for sharing Every with your friends. Join our referral program.