Mind Over Matter

AI in mental health care has tremendous potential. If it’s used responsibly.

As corporeal beings, we humans rely on our health, both physical and mental, and taking care of them is of utmost importance. With the advent of AI, we have new tools to help us in our mental health journeys. It’s important to remember, as AI ethics and innovation expert Songyee Yoon, founder of Principal Venture Partners, advises that while AI is a powerful tool, it’s not a magic wand. And like any tool, its effectiveness depends on how it’s used.

Promising Uses

In clinical settings, AI performs several functions, such as annotating conversations so practitioners can focus on the conversation with the patient rather than scribbling notes and missing key pieces of information. Services like Blueprint transcribe and write progress notes, and treatment plans.

The ability of AI platforms to sift through vast amounts of data and gather valuable insights augments the best care trained professionals provide. In such cases, it’s the best of both worlds.

Songyee Yoon

AI can read into “breathing patterns and voice patterns and some kind of other physiological data that there is. You can imagine how much data there is, but it’s really hard to discern what that means. But if you run it through the models and analyze it, it tells you a lot of things about your health, your emotions, your mental health, your behavior,” Yoon told Reboot.

Arduino’s Introvention, for instance, utilizes AI to detect small tremors that could be early signs of Parkinson’s disease but might otherwise go unnoticed until symptoms become more severe.

Yoon says when it comes to mental health care, there’s work in progress in many areas, with promising early results.

“There are companies still in stealth, a number of startups collecting all sorts of data and training models to predict the health of patients” whose conditions “we have not been able to detect or recognize or find the causality to apply more effective treatment.”

Down the Rabbit Hole

For responsible and ethical usage, it’s best to leave AI tools in the hands of the professionals. In increasing numbers, people have been turning to AI as a therapist. Even chatbots with safeguards built in can be manipulated into providing people with advice on self-harm and suicide methods, a Northeastern University study found.

Chatbots aren’t really designed to take on such responsibilities; they’re designed to prolong engagement, to keep people chatting.

“As we have seen in the social media cases, oftentimes these algorithms end up creating echo chambers. It can create a type of comforting message for those who are talking to them and not necessarily the kind of best advice, but something that people want to hear,” Yoon said. “That ends up strengthening or deepening the trust with a chatbot from the patient’s point of view and reinforces a certain way of thinking. I think there is a danger of going into the rabbit hole.”

Users can develop machine bias, thinking that any information from a chatbot comes from an objective final source of truth. The AI revolution came on so quickly that people didn’t really have a chance to wrap their minds around the possibilities before jumping in.

“We did not have enough time to get acquainted, recognize the limitations, and educate the general public to be more AI literate so that they understand that it’s an algorithm that comes with the limitations and to take their recommendations and whatever they say with a grain of salt,” she said.

Any use of AI in mental health needs supervision by real doctors, and health practitioners utilizing it need to be trained on, and bear in mind the potential and limitations of the tool.

Use it Wisely

The best use of technology, then, is not outsourcing important decisions to an unaccountable program. It’s using AI’s capabilities to do things we can’t easily, enhancing the work of mental health professionals, allowing them to do more groundbreaking work.

“This technology is at the end of day there to serve us and there to augment what we do,” Yoon said, “processing data, automating a lot of manual stuff so that we can carve out more time to spend building meaningful connections with a patient should be the objective.”

Protocols for safe care are in place for a reason, Yoon says, and they shouldn’t be set aside when a new technology comes along, no matter how exciting.

“I’m a technology optimist,” she said. “That’s why I work in this domain. But I do believe that it’s important to address it with a high level of care and responsibility.”

With the global market for AI in mental health forecast to grow 32% year over year through 2033, the stakes are high. AI is a powerful tool with a lot of potential for improving mental health care, Yoon says, but we have to use it wisely. And we have to remember that healing is the ultimate goal.