Akido Labs, a healthcare startup, is deploying Large Language Models (LLMs) to automate key aspects of medical appointments and enhance diagnostic processes. Their system, ScopeAI, is currently available at Akido’s Southern California clinics, promising faster access to specialist consultations. ScopeAI transcribes and analyzes patient-medical assistant conversations, generating potential diagnoses and treatment plans for physician review and approval.
Utilizing fine-tuned versions of Meta’s Llama and Anthropic’s Claude LLMs, ScopeAI tackles complex cognitive tasks, including patient history gathering and preliminary diagnosis formulation. While Akido Labs aims to improve healthcare efficiency and accessibility, the reliance on AI in critical medical decision-making has triggered discussions about potential risks and biases. Some experts are voicing concerns over automation bias, which could potentially influence physician judgment within Akido clinics.
Akido Labs is currently implementing ScopeAI in cardiology, endocrinology, and primary care settings, as well as through their street medicine program. Impressively, their street medicine team reports providing patients struggling with substance use disorders with access to crucial medications within 24 hours.
While Akido maintains that ScopeAI necessitates human oversight and is engineered to minimize bias, some observers worry about the transparency of AI’s role in patient care. Patients may not always fully comprehend the degree to which AI is shaping their medical journey, which could raise ethical considerations.