Dr. Tara Burra, Medical Director of Quality, Experience and Safety at CAMH, leads system-level efforts to improve patient safety, clinical quality, and care experience. In this interview, she discusses her path into psychiatry, current challenges, AI’s growing role in mental health, and its implications for patient-centered quality and safety.
Background and Motivation
Can you share a bit about your path into psychiatry, what initially drew you to the field, and what has kept you here?
I became interested in psychiatry during my clinical rotations in medical school. I had originally planned to pursue public health, but I found myself drawn to psychiatry because it allowed more time to sit with patients, understand their experiences and think deeply about how the brain shapes behaviour. These same elements have kept me in the field. I value the opportunity to make a meaningful, sustained impact on people’s care over time, and I remain fascinated by the brain’s complexity. Over the years, I have also become increasingly focused on the quality of psychiatric care. The field continues to face challenges related to access, consistent delivery of evidence-based treatments, and the ability to translate research into everyday clinical practice. These gaps in knowledge translations highlight the ongoing need for improvement in psychiatry.
Psychiatry Today
What do you see as the most pressing challenges in psychiatry right now, whether that be clinically, operational or system-wide?
Clinically, one of the most persistent difficulties is diagnosis. Unlike many other areas of medicine, psychiatry lacks laboratory tests, imaging, or reliable biomarkers that can confirm a diagnosis. Clinicians must rely on behavioural observations and patient-reported symptoms to guide assessment and treatment planning. This makes the field extremely nuanced and underscores the need for basic science research to better understand brain functioning. The brain’s complexity contributes significantly to the uncertainty when it comes to diagnoses that are present in psychiatric practice.
At the system level, underfunding continues to be a major barrier. Mental health and substance-use care have received a disproportionately small share in healthcare funding in Ontario and other territories. Although recent prioritization of these services has shown some improvement, this gap is still ongoing. This issue was compounded by the deinstitutionalization movement of the 1970s and 1980s that shifted mental health care away from large psychiatric asylums. Though this addressed real concerns, it was not accompanied by adequate investment in community-based services.
Operationally, data infrastructure poses significant challenges. Mental health care is delivered across diverse settings like community agencies, psychotherapy providers, rehabilitation programs, primary care and hospitals. However, these sectors operate on disconnected electronic health record systems. These fragmentations make it difficult to understand a patient’s history and coordinate care effectively. Even within hospitals, integrating measurement-based care into electronic health records has been difficult. This is because many systems were not designed with mental health workflows in mind and do not accommodate patient-reported outcomes or experience measures.
How have patient needs or their expectations changed over the past few years, especially post-pandemic?
Patient expectations have changed considerably, particularly regarding virtual care. Many patients expect the option of virtual visits, and some even request telephone-only appointments. Clinically, that can be challenging, as it is difficult to conduct a thorough mental status examination without seeing the patient. The pandemic accelerated a major transition toward virtual care. Previously, patients often had to travel to designated hubs connected to the Ontario Telemedicine Network. Now, a wide range of platforms allows people to connect from home, improving accessibility for many. Psychotherapy had also transformed significantly. The expansion of virtual psychotherapy, both video-based and telephone-based, has been one the most notable developments. Strong evidence shows that virtual psychotherapy is as effective as in-person care, helping broaden access and reduce barriers for individuals who may have struggled to engage with traditional models.
AI’s Presence in Mental Health (Quality & Safety)
From your perspective, how is AI already showing up in psychiatry?
AI is beginning to appear in psychiatry in several meaningful ways. A recent Grand Rounds session I attended highlighted multiple categories where AI is influencing care. In my own practice, the most direct exposure has been through ambient scribe software, which we have been testing in mental health settings to support documentation and reduce administrative burden.
There is also growing research on AI-enabled chatbots and digital tools designed to support psychiatric or psychotherapeutic interventions. Some of this work focuses on triage and scheduling, using AI to help route patients from primary care to the most appropriate clinician or service.
Patients themselves are also engaging with general-purpose AI tools, like ChatGPT or Claude, to ask clinical questions or seek life-coaching or therapy-like support. Although these platforms are not designed for clinical care, people are turning to them for guidance and sometimes bring those interactions into appointments.
On the research side, predictive modelling is a major area of interest. Teams are exploring whether AI can help forecast clinical outcomes more accurately, including risks such as future violence or other complex behaviours. A wide range of prediction tools are being tested to determine whether they can meaningfully support clinical decision-making.
Do you see AI as a complement to psychiatric care, fundamentally changing it, or landing somewhere in between?
I believe AI has strong potential to serve as a complementary tool in psychiatric care. As with any intervention, it must undergo rigorous testing to ensure safety, reliability, and clinical appropriateness. Its value has been demonstrated in areas with limited access to psychiatrists, including many northern and remote communities. In those settings, well-validated and screened AI tools could expand access to help address longstanding service gaps. It is important to once again note that though promise is high, safety and evidence must remain central. If these tools are thoroughly evaluated and proven safe, they could become a meaningful complement to the care that clinicians already provide.
Looking Ahead: AI, Quality and the Future of Care
If you imagine psychiatry 10-15 years from now, what role do you think AI will realistically play?
I don’t think anyone truly knows what the future will look like, and I don’t believe humans will be removed from the system any time soon. That said, there is a real potential for certain AI tools to outperform humans in specific areas. For example, early research suggests that some chatbots are perceived as more empathic and less judgmental than clinicians, which could be an advantage if study results hold up with this finding.
The key issue is whether these strengths are consistent, safe, and reliable. We are not near a point where mental health care could run on “autopilot.” Psychiatric presentations are far too complex for that. Unlike other fields, we cannot rely on a single MRI findings or lab value to make a diagnosis. Many patients have multiple co-occurring conditions and complicated social contexts. For example, for an older adult who struggles with depression and misuses alcohol, determining safe medication dosing requires integrating age, comorbidities, substance use, and psychosocial factors.
More advanced forms of AI, like agentic AI, may eventually handle more complex cases, but safe psychiatric practice currently still depends on nuanced, individualized judgement.
What gives you optimism about the intersection of technology and mental health?
I’m generally optimistic about AI because it has real potential to expand access to care. Many people could benefit from easier access to reliable information and evidence-based interventions which AI may be able to help with, closing some longstanding gaps in the current psychiatric framework. At the same time, there are important risks to consider, including environmental impacts and broader concerns about safety, equality, and responsible implementation. Any benefits must be balanced with thoughtful oversight.
The pace of change, with AI being implemented in psychiatry, also makes this an exciting but unpredictable moment. It reminds me of past technological shifts, like the introduction of paper, that fundamentally changed how information was shared. These transitions are hard to predict in real time. My hope is that we approach AI with due diligence. Hopefully we can harness its benefits while proactively mitigating the risks, so people ultimately receive better, safer and more accessible care.