Doctors dictate because speaking is dramatically faster than typing, and modern medicine demands an enormous volume of written documentation for every patient encounter. The average person speaks at roughly 150 words per minute, while most physicians type only about 21 words per minute. That speed gap, combined with the sheer amount of paperwork required for billing, legal protection, and patient safety, makes dictation one of the most practical tools in a doctor’s workflow.
The Speed Gap Is Massive
A 2025 study measuring clinicians’ actual keyboard and dictation speeds found a median typing rate of just 21.4 words per minute compared to a median dictation speed of 93 words per minute. That’s a 4.3x speedup for the same physician doing the same work. Even professional typists, who average 65 to 75 words per minute, can’t match ordinary speaking pace. For a doctor who needs to produce detailed notes on 20 or more patients a day, that difference translates to hours saved.
This matters because primary care physicians already spend roughly 5.9 hours of an 11.4-hour workday working in electronic health records. That breaks down to about 4.5 hours during clinic time and another 1.4 hours after the clinic closes, often late at night. Research from the Annals of Family Medicine found that physicians spend nearly 2 hours on electronic records and desk work for every 1 hour of direct patient care. Dictation is one of the few tools that compresses the documentation side of that equation without cutting corners on detail.
Medicare and Insurance Require Detailed Notes
A major reason doctors produce so much documentation in the first place is that insurance reimbursement depends on it. The Centers for Medicare and Medicaid Services requires sufficient documentation to verify that every service billed was actually provided, was medically necessary, and matched the level of care charged. If the notes are incomplete or lack detail, the claim can be denied outright.
CMS auditors specifically flag incomplete progress notes, missing signatures, undated entries, and records that don’t explain why a test was ordered or a treatment was chosen. Office visits, initial hospital encounters, and follow-up hospital visits rank among the most common billing errors, often because the documentation didn’t support the level of service billed. In practical terms, a 10-minute office visit can require several paragraphs of structured documentation covering the patient’s history, the exam findings, the clinical reasoning, and the treatment plan. Dictation lets physicians narrate all of that in close to real time rather than hunting and pecking through a keyboard afterward.
Legal Protection Depends on the Record
Medical records are legal documents. In malpractice cases, inaccurate, incomplete, or careless records undermine a physician’s defense and make a plaintiff’s attorney more likely to pursue the case. In one notable example, a court found a physician partially liable when a patient left against medical advice because the conversation about risks wasn’t documented, even though the patient had signed an AMA form. The signed form alone wasn’t enough without a note describing the discussion.
Some courts go even further. If a record appears to have been altered after the fact, the burden of proof can shift entirely: instead of the patient proving the doctor caused harm, the doctor must prove they didn’t. Dictation helps here because it encourages physicians to create thorough, contemporaneous notes while the encounter is fresh. Speaking naturally tends to produce more complete narratives than typing shorthand that gets cleaned up later.
How Dictation Has Changed Over Time
For most of the 20th century, dictation meant a doctor speaking into a tape recorder or dictaphone. A human transcriptionist would then listen to the recording and type it into the medical chart, sometimes hours or days later. This was already faster than having physicians write or type everything themselves, but it introduced delays and the cost of transcription staff.
Computers in the 1980s and 1990s replaced tape recorders with digital recordings, improving audio quality but keeping the same basic workflow. The real shift came in the 2000s, when voice recognition software matured enough to convert speech to text in real time. Physicians could dictate directly into electronic health record systems, bypassing transcriptionists entirely. Today, the latest iteration is ambient AI scribes: software that listens to the entire doctor-patient conversation and automatically generates a structured clinical note afterward, with no dictation step required at all.
Accuracy Tradeoffs Are Real
Dictation isn’t perfect. Speech recognition error rates vary widely depending on the setting. In controlled dictation environments where a physician speaks clearly into a microphone, word error rates can be as low as about 9%. But in conversational or multi-speaker scenarios, like a busy exam room, error rates can climb above 50%.
The nature of errors differs between dictation and typing too. One study found that speech recognition produced 142 typographical errors compared to 71 for manual keyboard entry, and 29 errors with potential for patient harm versus 11 from typing. However, a separate study found that dictation left fewer uncorrected errors on average (1.5 per note versus 2.9 for typing), likely because physicians reviewing dictated text catch obvious speech-to-text mistakes more readily than they catch their own typos. The correction workload was also lighter: dictated notes averaged 4.1 corrected errors per document compared to 33.9 for typed notes, suggesting that typing required far more editing to reach a final version.
Most physicians who dictate develop a habit of reviewing the generated text before signing off, catching misrecognized drug names or garbled phrasing. The tradeoff is generally accepted because the time savings outweigh the review effort.
Impact on the Patient Visit
One persistent concern about electronic records is that doctors spend too much time looking at screens instead of patients. Research has confirmed that many patients feel exam room computers reduce interpersonal contact with their physician. Dictation, and especially ambient AI tools, can help with this by removing the need to type during the visit itself. A doctor who dictates after the patient leaves, or who uses an ambient scribe that records the conversation passively, can maintain eye contact and focus on listening throughout the appointment.
That said, the relationship between screen time and patient satisfaction is more nuanced than it first appears. A pilot study in ophthalmology found that the amount of time a physician spent looking at the computer versus the patient didn’t significantly affect satisfaction scores. Most patients in that study were comfortable with EHR use during visits and even felt the technology made their care more effective. Existing trust and an established relationship with a doctor likely matter more than whether the doctor types, dictates, or uses a scribe.
The Shift to Ambient AI
The newest chapter in medical dictation barely involves dictation at all. Ambient AI scribes record the natural conversation between doctor and patient, then use artificial intelligence to generate a complete clinical note. A large study published in JAMA Network Open found that 44.6% of physicians offered the tool adopted it. Those who used it saw about 0.80 more patient encounters per week than nonadopters, with no increase in insurance claim denials. The tool is associated with reduced documentation time and lower rates of burnout, which helps explain why adoption is climbing quickly.
For physicians, the appeal is straightforward: the technology handles the most time-consuming part of the job while letting them stay fully present with patients. For patients, it means a doctor who listens and talks to them rather than a doctor who types.