
Artificial Intelligence in the Doctor’s Office
By Sarabeth Lowe, MPH Ms. Lowe is a Communication Specialist at the University of Delaware Disaster Research Center. Welcome to the second edition of Health Literacy Highlights, a new NHR column where I explore topics related to finding, understanding, and utilizing health information. This column is meant to empower you with the skills you need to apply to what you already know and how to use it to maintain and protect your overall health and well-being. In this installment of Health Literacy Highlights, I will be discussing the use of ambient artificial intelligence (AI) in healthcare. If you’ve visited your doctor recently, this scenario might sound familiar: You call the doctor’s office only to find that the next available appointment, whether its with a specialist or a primary care physician, is months out. You schedule the appointment, and you wait for the day to arrive. You wait for your name to be called in the waiting room. Finally, you’re waiting for the doctor in the exam room. When you finally meet with the doctor, the conversation is brief and feels rushed. To make matters worse, their eyes are mostly glued to their computer screen, making the exchange feel impersonal and brusque. Even so, you know that you should not take it personally. After all, the doctor needs to move along to their next patient. Still, you leave feeling a little slighted, maybe even snubbed, after the interaction. Of course, this is not the desired scenario for the doctor, either. Its no wonder that patient and physician burnout have become a serious issue in public health.1–6 According to the Association of American Medical Colleges, the United States will face a physician shortage of up to 86,000 physicians by 2036.7,8 Luckily, there is help—and hope—on the horizon. AI is transforming the US healthcare system.6,9–12 As these tools continue to evolve and improve, many healthcare providers and experts agree that this innovative technology will play a critical role in shaping a more positive future. At the same time, enthusiasm for this technology is growing because they feel confident that leveraging these tools is finally a reality.13–18 Numerous tools have come onto the scene, including AI-powered chatbots, automated appointment scheduling, and computer-aided diagnosis algorithms. One, however, has emerged as a particularly effective tool for addressing patient-physician communication and burnout: ambient AI. What Is Ambient AI? Ambient AI, which I will refer to as AAI, is a subset of AI that focuses on creating intelligent systems—machines capable of recognizing and responding to their environment—that can solve problems in real-time and at scale.19,20 It uses ambient listening technology (ALT) and a combination of different AI-powered tools, including machine learning, data analytics, and natural language processing, to capture, process, and interpret conversations in near-real time. You can think of AAI as having a medical scribe in the room that never needs training, never burns out, and never interrupts. In clinical settings, healthcare providers use a recording device to “listen” in during their appointments, with the patient’s consent. AAI then analyzes this doctor–patient dialogue and converts it into structured medical documentation.21 Unlike other types of AI, where the user must ask a specific question or provide explicit context or details, ALT does not require a particular prompt or manual input. As the term “ambient” describes, this technology is non-intrusive and exists in the background. It only activates when the device perceives a context cue, such as a patient entering an exam room or an ID badge getting scanned. In short, AAI is designed to work behind the scenes without human intervention and be responsive, proactive, and intuitive. How It Works and What to Expect AAI is more than just a recording technology. It does not just hear; it listens. AAI captures clinical context and transforms it into clean, usable medical documentation. It is able to identify symptoms, medications, diagnoses, and treatment plans. AAI accounts for all of this information and context—extracted from the raw, unedited dialogue between you and your doctor—and turns it into a high-quality, comprehensive note that fits neatly into your electronic health record (EHR).20,21 It is even able to recognize the nuances of conversation, such as pauses, voice inflection, and words per minute, and include those details in the digital record. According to Freed, a technology company behind a popular medical AI scribe tool that more than 20,000 physicians have used, here is what happens behind the scenes: Recording. A context cue will wake the ALT device, such as a tablet, smartphone, or ceiling microphone, which will be passively capturing audio from the room. Speech-to-text conversion. During the appointment, the ALT system will use automatic speech recognition to convert the conversation into text. This technology is able to account for multiple speakers. Speaker ID and context detection. ALT will use advanced diarization (a fancy term for “who said what”) to identify the speakers (eg, patient, doctor, nurse, etc.).21 It also identifies clinical context, including symptoms, conditions, medications, and treatment plans, during this time. Structuring the note. After the conversation is recorded and transcribed, the system will organize it into a structured format, such as a Subjective, Objective, Assessment, and Plan (SOAP) note. Instead of just transcribing the dialogue, ALT is able to transform the information into a usable clinical note. Clinician review and handoff. Before this documentation is finalized and added to your EHR, the note is handed back to your provider, who will review the note, make any needed edits, and sign off to confirm its accuracy. AAI is able to complete these steps within minutes or seconds, depending on the system.21 This entire process, from the start of the appointment to writing up clinical documentation and submitting it to your Electronic Health Record (EHR), is time- and labor-intensive for healthcare providers and requires meticulous attention to detail. Research indicates that primary care physicians spend approximately 36 minutes on EHR documentation for each patient. All of this extra time spent on clerical and administrative tasks, which can add up to hours of additional