���۴�ý

Trying to understand AI and mental health requires a different perspective. Imagine you’ve just come out of a deep freeze, and you’re trying to get up to speed on 21st century America. Well, speed becomes the key word. The alacrity with which artificial intelligence has moved into business, education, healthcare and almost every other area of life is incredible.

But is AI in our minds too? Can it be? Should it be?

Ruminations about the ethical use of AI aside, the tool does hold promise with regard to positive mental health benefits. Consider the following.

Apps aimed at offering mental health support are proliferating. In this way, AI offers a deep advantage. It’s accessible during the same hours that mental healthcare is needed: 24/7.

“If people have access to a computer or a phone, then they can use some of these tools and that’s kind of like 24/7 support,” Luster says. “Of course, AI is never going to replace a therapist, because you’re talking about the human quotient in that. However, as a mental health resource, it becomes available on a 24-hour timeline, which is pretty cool.”

AI is showing some promise in helping detect mental health problems. This might be as innocuous as AI detecting via your smartwatch that your physical activity has gone from high levels to zero, possibly indicating depression. 

Or it could simply mean you’re on an impossible work deadline. Either way, a chatbot prompt to see if you’re OK or to remind you to move isn’t a bad thing.

A review of the literature on AI and mental health, published in , reveals that through use of technologies such as wearables or smartphone applications that offer real-time emotion data, AI can provide accurate detection (perhaps in some cases even surpassing human detection). 

In other cases, AI can help people connect patterns that empower them to take charge of their health. For example, AI tracking tools that provide users with insights to their emotional patterns over time can help individuals understand their triggers. This, in turn, can help them develop healthier responses to those triggers.

Another possible scenario could use algorithms to analyze MRI images, making early detection of difficult-to-diagnose diseases such as Alzheimer’s or Parkinson’s easier — thus making it possible to intervene earlier. 

According to that same Frontiers in Digital Health study cited above, the digital triage capabilities of emerging AI-powered apps and programs may not stop at detecting mental health concerns. They might also be programmed to intervene.

For example, as AI analyzes a person’s answers related to mood, stress, energy and sleep patterns, a chatbot might recommend behavioral modifications such as physical activity, meditation and relaxation techniques. If the chatbot detects an immediate safety concern, it could promptly notify the patient’s healthcare provider.

If all this sounds a little too Big Brother-ish, consider that AI is showing enormous potential with certain, targeted populations.

For example, AI-powered therapeutic games and virtual reality experiences offer immersive environments that help those who experience emotional dysregulation practice their emotional regulation skills.

AI analysis of physiological and behavioral inputs from wearable technology and mobile devices is also showing promise for early detection of depression and bipolar disorder.

And in elderly populations, AI-enabled virtual companions can engage older adults in cognitive exercises and reminiscence therapy — addressing feelings of isolation and even cognitive decline.

As with any new frontier, especially in today’s rush-to-market world, AI has its downsides. 

“Whether you’re in business or you’re someone who’s trying to work through issues, I think AI requires mindful, thoughtful implementation,” Luster says. “How are you using it? Are you being strategic while you’re guarding against the risk of things with it?”

Here are some of the potential downsides.

An obvious drawback to the use of AI-generated applications and responses for something as deeply nuanced and internal as mental health is that it doesn’t substitute for human interaction. AI will never offer the compassion and human contact that come from sitting with a person.

Put a different way: “AI won’t release oxytocin,” Luster says. “It doesn’t have that human element.”

Proponents of AI’s mental health applications might say, “Well, it could trigger some of that.” But AI will never be the same as sitting across from a real human who can interpret facial reactions, tone and body mannerisms and then respond with empathy, gentility, reason and, if needed, a challenge.

Played out to an even deeper conclusion, AI may lead some users into a world of social isolation, with chatbots for friends and screens offering their only view into the world. The U.S. Surgeon General recently released an advisory calling isolation an epidemic, with the mortality consequences of social disconnection being roughly equivalent to smoking 15 cigarettes a day.

Luster cautions that if people rely on chatbots, app-driven interactions and the virtual world rather than human connection, then one of our core human needs — the need to belong — will go unmet in the real world. “AI will only go so far,” he says.

Innovative solutions to mental healthcare delivery are needed in a system that is currently overloaded. However, the speedy release of AI-enabled models has some sounding the alarm. Fast doesn’t mean factual. Accuracy with mental health diagnosis and intervention is critical. 

In a stroke of irony, the very AI-enabled transcript from the videoconferencing program used for the interview with Luster issued this warning prominently over the digitally generated transcript: “AI generated content may be incorrect.” So, there’s that.

A study published last year in the Journal of Medicine, Surgery, and Public Health suggests that some AI modelers are responding to the demand to approach mental health needs with accuracy. “Developing AI models that are interpretable and can provide explanations for their recommendations is a growing trend. This promotes transparency and allows clinicians and patients better to understand the reasoning behind AI-generated insights and decisions,” the study concludes.

But what about far less controlled settings, like the internet, for example?  

A points out that the application of generative AI to mental health subjects has a dark side. One study found that “AI tools have been providing users with harmful content surrounding eating disorders around 41% of the time.”

These are the kinds of issues that must be ironed out. But in the short term, Luster is as curious as anyone else. 

As for that acquaintance who was going to give his girlfriend a piece of his mind? AI actually changed his mind. When AI held up a mirror to the potential injury his words might cause to someone he cares about, he changed course.

“He later reflected on that, and when he went to visit his girlfriend that night, he changed his whole approach,” Luster says. “He enhanced his own emotional disposition and understanding more dimensionally by considering both parties’ feelings in this case.” 

Luster pauses before adding the most important part. “He said they had a wonderful evening.”

AI and mental health represent just one category of exploration. For those looking to enhance their AI knowledge in general, ���۴�ý recently launched an undergraduate elective course that offers hands-on experience in an instructor-led environment. 

Known as Generative AI in Everyday Life, the course also builds durable skills like creativity and adaptability while teaching students how to effectively and ethically use AI.

Intrigued? Request additional information!