Waiting times in psychiatry are increasing, and documentation requirements are placing growing pressure on clinicians’ daily work. At the same time, artificial intelligence is gaining ground across healthcare. The question is no longer whether the technology will play a role, but how it can be implemented responsibly and meaningfully in clinical practice.
One of the voices engaging with this development both curiously and critically is consultant psychiatrist Mikkel Rasmussen. A former medical director, senior consultant, political actor, and now practising psychiatrist, he brings experience as a clinician, system critic, and interdisciplinary collaborator. In recent months, he has worked closely with the development and testing of AI tools designed specifically for psychiatry.
“AI is not a replacement for the psychiatric gaze,” Rasmussen explains, adding, “It is a tool that can free up time, not judgement.”
AI as Support on Professional Terms
Mikkel Rasmussen sees clear potential in AI, particularly as support for structuring, questionnaires, transcriptions, and extensive assessment processes. According to him, clinical conversations are demanding in themselves, and the documentation that follows can be equally time-consuming.
“Here, AI can improve efficiency without compromising professional standards. But it requires the psychiatrist to remain in control,” Rasmussen elaborates.
For him, this is not about automating clinical assessments, but about intelligently supporting workflows. He highlights three areas where AI is already demonstrating tangible value in practice.
First, the structuring of lengthy consultations. AI can transform a complex dialogue into a clear draft clinical note. This creates greater calm during the patient encounter and allows for stronger presence without losing important details.
Second, questionnaires, rating scales, and semi-structured interviews. Here, the technology can support existing clinical protocols and contribute to more consistent documentation.
Third, documentation under pressure. When assessment interviews last between one and a half and two hours, AI can significantly reduce the subsequent administrative burden and free up clinical capacity.
Where AI Cannot – and Should Not – Replace the Clinician
At the same time, Mikkel Rasmussen is clear about the boundaries. AI may serve as a tool, but it must never become the decision-maker.
“One must always maintain a healthy scepticism and AI cannot replace the physical meeting and the pattern recognition that an experienced psychiatrist ultimately has.”
AI can lose nuance and tends to generalise. What matters most, he emphasises, is not simply the note itself, but the review and verification of it.
Clinical judgement remains central, particularly in relation to risk assessment, complex differential diagnoses, relational understanding, and interpretation of mood, ambivalence, and motivation.
“AI can do a great deal, but it cannot pick up on the nuances of the clinical encounter,” Rasmussen adds.
It is precisely here that implementation becomes critical. Technology must be designed and applied on the profession’s terms, not the other way around.
How Mikkel Rasmussen Uses AI in Clinical Practice
Mikkel Rasmussen has personally tested several digital solutions aimed at easing the documentation workload. His experience was that many tools functioned in isolation and produced fragmented results, with a risk of becoming overly generalised in their output.
He was therefore positively surprised when he tested AI in a more integrated format.
“A consolidated tool was more interesting – and the potential became clear quickly, especially when AI is continuously trained in clinical language, terminology and the structure of different types of notes and records,” he explains.
Today, he mostly uses AI in connection with assessments. The technology helps him structure large amounts of data, transcribe consultations, and generate an initial draft clinical note. At the same time, it contributes to more consistent and transparent documentation.
In fact, Mikkel Rasmussen has a clear goal: less administrative friction and more time in the actual conversation and encounter with the patient.
“I always trust my professional judgement the most. AI is a supplement, not a replacement,” Rasmussen states.
The Future Vision: A Stronger Partnership Between Human and Machine
According to Mikkel Rasmussen, the ambition is not a more technical psychiatry, but a more human one. If AI can manage repetitive and structural tasks, the clinician can focus on the relationship, which forms the very foundation of psychiatric treatment.
He identifies three potential areas of development, provided AI is implemented thoughtfully.
First, better preparation before consultations. Patients can share relevant information from home. A short AI-supported interview prior to the first appointment could provide the foundation for a more focused and qualified start to the treatment process.
Second, improved continuity in long-term treatment. Ongoing symptom tracking can be translated into structured overviews of development and patterns. This may strengthen dialogue between patient and clinician.
Third, more time for the relational core. When documentation occupies less space, presence can occupy more.
A Professional Call to Colleagues
Mikkel Rasmussen encourages colleagues to approach AI with openness and critical reflection, and to test the technology in practice. Not with uncritical enthusiasm, but with professional curiosity.
“When AI becomes too general and loses the patient’s unique essence. The examples, the small concrete statements, are often decisive in a long-term course of treatment.”
Balance is key. Technology must support clinical practice without erasing individuality.
“One must maintain a healthy scepticism and verify as in so many other areas, but it will save you time in the long run,” he says.