Key takeaway: Google DeepMind introduces its research program for AI co-clinicians, investigating how artificial intelligence can support doctors as genuine team partners. The concept of “triadic care” envisions AI systems caring for patients under physician supervision to improve the quality and availability of healthcare.
Google DeepMind is introducing a new research program investigating how artificial intelligence can support doctors. The initiative aims to demonstrate how AI systems can function as genuine team partners in patient care while physicians retain clinical control.
Health systems worldwide are striving for better treatment outcomes, reduced costs, and improved experience for patients and clinicians alike. However, progress is hindered by a global shortage of clinical professionals. The World Health Organization forecasts a deficit of over 10 million healthcare workers by 2030.
While AI is frequently seen as a solution to this gap, it has not yet met all the requirements of clinicians and patients. Google DeepMind today introduces its research initiative for AI co-clinicians to investigate how artificial intelligence can enhance physicians’ expertise and improve the quality of patient care.
Google DeepMind’s medical AI work has evolved from expert-level performance on medical knowledge exams with MedPaLM to matching physician capabilities in text-based simulated consultations with AMIE. The company also has a long track record in researching collaboration between clinicians and AI.
The central hypothesis is the concept of “triadic care”: AI agents support patients throughout their entire medical care under the supervision and clinical authority of their physicians. Medicine has always been teamwork, and AI systems can expand this team by augmenting the capabilities of clinicians while preserving their judgment and control.
Google DeepMind has developed and evaluated an AI co-clinician in both clinician-oriented and patient-oriented settings. The key to improving healthcare quality, cost efficiency, availability, and overall experience lies in considering both perspectives. Further research is needed to make medical AI systems more trustworthy and genuinely useful for clinicians in caring for patients. Ultimately, the goal is to equip clinicians with reliable AI co-clinicians. For a physician, a tool is only valuable if it is trustworthy and based on solid facts.