Can AI tools help solve our mental health crisis or is it too risky?

Struggles with mental health are on the rise across the world.

Last year the World Health Organization said one in eight people globally live with a mental health condition. That's 970 million people - up from 900 million the year before.

Support for these conditions is growing. Last year the global mental health market was valued at over $450 billion. It's predicted to be worth $623 billion by 2032.

Technology is playing an increasing role, with tens of thousands of apps available, increasingly using artificial intelligence to tailor responses to those seeking help.

But should we be worried about this? There are concerns about how much of our personal information is being shared with private companies, unbeknown to us, threatening our privacy. 

There are also worries that relying on AI tools may see serious impending crises go undiagnosed and a wave of misdiagnosis. 

Clinical psychologist Professor Bethany Teachman, who runs the Program for Anxiety, Cognition, and Treatment (PACT) lab at the University of Virginia, says AI tools have a host of potential benefits but insists that careful regulation is key to their efficacy.

She told CGTN's Robyn Dwyer: "We have some nice evidence that the digital tools that are based on research and use principles that we know can really help deal with anxiety and depression, particularly cognitive behavior therapy principles, can work just as well in some cases as in-person care."

Teachman - an expert on anxiety - said the key to ensuring that digital apps don't miss vital signs of distress is working with stepped care models, "the idea being that if you have maybe less serious problems or you have problems that are not likely to lead you to have an acute crisis, then it may be appropriate to use these devices."

She said that dealing with issues that are more complex or serious should usually require in-person help but also believes blended care could be transformative. 

"From the standpoint of say a treatment for social anxiety disorder, I might need to see somebody 16 times. That's a lot of time. If we use these digital tools, maybe I can see the person eight times and the digital tool can do some of that work as well."

Teachman believes that with the shortage of mental health help amid a surge in referrals in the wake of the COVID pandemic, "it's really important that we think about these innovative care delivery models to increase access."

Her own research focuses on how biased thinking contributes to anxiety disorders and Teachman says there are clearly questions about the ethics of using AI chatbots. "We must make sure that they're safe and appropriately regulated…..(but) if we can work on the safety issues, they could really increase access to care."

Teachman believes AI could also be helpful in training the mental health care workforce and improve the personalisation of treatment. 

On the risks to patients' confidentiality, Teachman warned: "It's important that people are really monitoring that there aren't companies or governments who are inappropriately sacrificing our privacy because obviously these topics are incredibly sensitive. It's absolutely important that we think about regulations in that space and being really careful in what we're doing."

Using offline versions of digital tools is one answer. "In the work that we're doing, for example, we never upload any information about individuals into the open AI platforms - never - and we have no intent in doing that."

Last year, five percent of adults globally had depression and 4 percent suffered from anxiety. 

Teachman insists the challenges of coping with almost a billion people living with these conditions means technology has to be part of the answer.