Artificial Intelligence inheriting societal biases may hinder mental healthcare: Experts | Bengaluru News

Bengaluru: Artificial Intelligence (AI) is increasingly being used in healthcare, promising efficiency and accessibility. However, when it comes to mental health, AI may be deepening existing gender disparities rather than resolving them. Experts warn that AI models trained on biased data risk misdiagnosing women and providing inadequate treatment, perpetuating long-standing inequities in mental healthcare.
At the 10th World Congress of Women’s Mental Health on Thursday, researchers and clinicians discussed the growing concern over AI’s role in diagnosing and treating mental health conditions.
Dr Ana Barajas, psychologist and associate professor at Autonomous University of Barcelona, Spain, highlighted a key issue: “Gender bias in AI is worsening disparities in mental health diagnosis, treatment, and outcomes for women. AI models often rely on data that under-represents women’s mental health conditions, making them less sensitive to their needs.”
AI’s reliance on historical data means it can inherit and reinforce biases present in traditional healthcare. One example is the way gender differences in language patterns are often overlooked. Since AI systems analyse text and speech to assess mental health, failing to recognise these differences can lead to inaccurate diagnoses.
Dr Raquel Iniesta, an expert in machine learning and ethics at King’s College London, explained that healthcare has historically been biased toward male symptoms and AI is no exception. “We need to ensure AI models are built on representative data, rigorously tested, and free from discrimination before they’re widely used.”
Fixing systemic biases
Beyond gender bias, broader societal disparities are also at play. Dr Jyotishman Pathak, professor of medical informatics and psychiatry at Cornell University, US, said: “Our mental healthcare system is already deeply biased, by race, socio-economic status, and access to insurance. AI trained on this flawed data inherits and amplifies these disparities. Instead of just fixing AI downstream, we must address systemic biases upstream to create fairer, more inclusive care for all.”
Another concern is how AI affects the human aspect of mental healthcare. Dr Pathak noted that in a focus group with 500 patients and stakeholders in New York, a common fear emerged: Will AI weaken the therapeutic relationship? “Mental health already carries stigma — especially during pregnancy and postpartum. Will AI make patients even more hesitant to share symptoms? We need stronger safeguards now.”
Despite these challenges, experts agree AI has potential benefits. AI-driven mental health tools can increase accessibility and reduce stigma, allowing people to seek support anonymously and at any time. However, these tools have significant limitations. “They lack human empathy and struggle with cultural nuances,” said Dr Barajas.
Tackling postpartum psychosis
Psychiatrists have raised an alarm over the increasing incidence of postpartum psychosis in India, now affecting two in every 1,000 births. The severe mental health condition can emerge suddenly within days or weeks after childbirth, leading to symptoms such as extreme mood swings, confusion, and psychotic episodes.
Dr Sonia Parial, a seasoned psychiatrist, told TOI, “Postpartum psychosis is a severe psychiatric emergency that can escalate rapidly, often within days to weeks of childbirth. In India and globally, it has a high risk of suicidal infanticide, yet remains under-recognised. Women with bipolar disorder face a greater risk, especially if they stop medication during pregnancy. Symptoms range from extreme irritability to catatonia, often focused on infant-related delusions.”
Dr Ruksheda Syed, president, Bombay Psychiatric Society, highlighted the importance of mother-baby units (MBU) to tackle postpartum psychosis. “In many cases, women with postpartum mental health issues end up with gynaecologists instead of psychiatrists, as hospitals lack psychiatry departments. An MBU would solve grassroots problems.”