AI is promising to revolutionise how we diagnose mental illness
The last big breakthrough in treating depression was in the 1980s. That’s when Prozac, the first SSRI antidepressant, was released. This drug, along with its later copycats, quickly took over the world, and hundreds of millions of people have now taken this type of medication. But even though three-quarters of people say the pills helped them feel better, they don’t work for everyone. With rates of depression rising – and the lack of major therapeutic advances since the advent of SSRIs – it seems almost inevitable that some will hope that AI could become the next big thing in psychiatry.
The warnings are well known. Chatbots are only as good as the data they are trained on, have our own biases, and are prone to errors described by some as “hallucinations.” This week we have news of a study that found that some of the best-known AI models fail to give adequate advice for 60% of women’s health queries.
But AI could finally bring essential objectivity to the delicate question of diagnosis. Currently, depression is diagnosed through a list of vague and imprecise symptoms. By using AI to analyze subtle physical signs like our facial expressions and the cadence of our voices, psychiatry is finally beginning to create the more defined biomarkers that the field has long awaited.
“
AI could finally bring essential objectivity to the delicate question of diagnosis
“
The implications could be profound. In the absence of new biomolecular advances, much progress has recently been made in understanding the role that factors such as personal relationships and access to nature play in preventing depression. A recent review found the strongest evidence to date that exercise is as effective as antidepressants or cognitive behavioral therapy for treating depression, although it is not clear why or who is best suited for this treatment.
If AI can finally help us answer the question of who responds best to which treatments, it could change the lives of millions of people – provided developers guard against AI’s many flaws from the start. No one wants their psychiatrist to have hallucinations.
