AI Scribe Hub

As part of our work to help doctors reduce administrative burden, Doctors Manitoba is sharing resources to make it easier to integrate Artificial Intelligence (AI) scribes into your practice. 

In our 2025 annual physician survey, we found only 6% of physicians are using AI Scribes, but nearly 60% are interested in trying them out. Many physicians using AI scribes in Manitoba report saving 30 – 60 minutes or more per day. Many who have not yet started say they have concerns surrounding uncertainty about privacy and other rules, are not sure where to start, or feel overwhelmed with different vendor options.

This AI Scribe Hub will help you find current information and resources to help you adopt this promising technology into your practice. 

What is an AI Scribe?

AI scribes use artificial intelligence and natural language processing to help doctors with transcribing and summarizing patient interactions in real time. They have the potential to reduce administrative burden by automating repetitive tasks such as documenting patient visits, updating medical records, and summarizing treatment plans.

An important distinction between AI scribes from other AI applications in medicine is the role they play. AI scribes help with documentation, but they do not play a role in diagnosing or treating patients. 

AI scribes are relatively new. It’s important to understand what AI scribes are and how they could benefit your practice, as well as the risks and limitations to be aware of as a physician.

Benefits and Risks

Physicians report several benefits of using an AI scribe. These include:

🕒 Saving time on documentation, recognizing most physicians spend one to two hours per day on charting.

📃 Improving patient-physician interaction by reducing note-taking distractions, allowing physicians to listen more attentively and build a stronger relationship with each patient. 

Improved quality and accuracy of clinical notes, with AI scribes able to pull in complex medical information.

However, using AI scribes comes with some risks, which you must be aware of and address as a physician. This includes:

  • Privacy: As an extension of your obligation to protect and secure patient’s personal health information.
  • Accuracy: You are responsible for your documentation, whether it’s drafted by you, another human, or an AI scribe. As AI has to potential to inaccurate information, you must review, edit and correct AI drafts before signing off. 
  • Consent: At this stage, regulators require physicians to obtain patient consent about using technology to help with your documentation. 

➡️ You can learn more about this in a separate article here.

Other FAQs

How does it differ from a dictation software?

Dictation software converts speech to text, whereas AI scribes summarize the content of the encounter into a note. In some cases, it can even apply it to appropriate fields of your EMR.

Why use an AI scribe?

Many physicians in Manitoba spend over 10 hours a week on administrative work. Charting, while a necessary and important task, takes away from physician-patient contact time. AI scribes can be used as a tool to help optimize your efficiencies in your clinical days and allow you to be more present with your patients.

What are some potential risks?

Like all technology, there are limitations and potential for error. Physicians must be cautious in reviewing all notes generated by their AI software. Notes may be incorrect or biased. Physicians are responsible for the accuracy of their patient records. As long as you are diligent with your responsibilities, the risks should be negligible.

Do I need to obtain consent?

Yes, you need to obtain consent and document it in the patient’s chart. See our article AI Scribes in Your Practice: Your Responsibilities to read more on your consent obligations.

Can my administrative staff obtain consent on my behalf?

We don’t recommend your administrative staff obtain consent from your patient. Information can be provided by your clinic staff to help inform the patient on AI, however, the consent should be obtained by the physician. 

Do they come in other languages?

Most Canadian based software recognizes both English and French. Many programs have even more language capabilities. It is important to review what your chosen product offers to ensure it supports your clinical needs. 

Common AI Terminology
  • Artificial Intelligence: The ability of a digital computer or computer-controlled robot to perform tasks commonly associated with intelligent beings. This includes learning, reading, writing, creating, and analyzing. The American Medical Association uses the term ​“augmented intelligence” instead to focus on how AI enhances human intelligence rather than replaces it.
  • Generative AI: Deep-learning models that can generate high-quality text, images, and other content based on the data they were trained on. Generative AI models learn the patterns and structure of their input training data and then generate new data that has similar characteristics. 
  • Large Language Model (LLM): A large language model (LLM) is a type of artificial intelligence (AI) program that can recognize and generate text, among other tasks. LLMs are trained on huge sets of data — hence the name large.” In other words, an LLM is a computer program that has been fed enough examples to be able to recognize and interpret human language or other types of complex data.
  • Natural Language Processing (NLP): Deals with how computers understand, process, and manipulate human languages. It can involve things like interpreting the semantic meaning of language, translating between human languages, or recognizing patterns in human languages. It makes use of statistical methods, machine learning, neural networks and text mining.
  • Automatic Speech Recognition (ASR): A capability which enables a program to process human speech into a written format. While it’s commonly confused with voice recognition, speech recognition focuses on the translation of speech from a verbal format to a text one whereas voice recognition just seeks to identify an individual user’s voice.
  • ChatGPT: A well-known example of a generative AI tool using LLM proprietary software from OpenAI, allowing users to ask questions, provide context and refine results in a conversational format.