I firmly believe that human expertise in transcribing is essential for ensuring effective communication. Moreover, I am convinced that professional transcribers can maintain job security by being adaptable and committed to continuous improvement.

When it comes to the impact of artificial intelligence (AI) on the real-time transcribing profession, I believe that this isn’t an either/or situation. While AI has its advantages in certain situations and can help save costs, there will always be a need for professional transcribers, especially where ADA compliance is required.

Deaf and hard-of-hearing individuals can, and will, request different accommodations for different settings.

Our best TypeWell transcribers excel in professional and educational settings, outperforming automated captions without question. One of my colleagues recently expressed skepticism about the stability of AI for live events. Based on her own experience, which is extensive: “AI is not suitable for professional use. It just isn’t.”

This distinction between AI and human skills isn’t just about accuracy and precision; it’s about context, nuance, decision-making, and problem-solving. Transcribers rely on their skills and expertise, often without conscious effort. This level of real-time decision-making and customization results in more effective communication.

Let me share some real-world examples to illustrate what professional transcribers are able to do with relative ease:

• When people talk over each other, a transcriber skillfully discerns which comments are relevant to the conversation and the reader(s) they are supporting. In some situations, they may remind the group or class to speak one at a time.
   
• If people suddenly begin talking over each other—with passion!—a provider can quickly understand what triggered this outburst and convey it to the reader efficiently. A TypeWell transcriber would customarily summarize the situation using a transcriber comment (enclosed in square brackets), to make sure the reader also has time to weigh in!
   
• A deaf student might ask a transcriber not to transcribe their own comments during classes; however, they want an exception to this rule during test review sessions. A human transcriber can easily honor this preference.
   
• When a webinar host asks a transcriber to use speaker identifiers selectively, or to redact sensitive or confidential information from the saved transcript, a professional service provider will accommodate this request.

To facilitate two-way communication, educational transcribers are trained to voice for readers, confirm their own understanding (when necessary), and advocate for clear communication.

Human transcribers/interpreters are skilled at “reading the room,” and they frequently interpret more than just words. That is why we sometimes refer to this work as “text interpreting” rather than simply “transcribing.” They can perceive the emotion in a speaker’s tone of voice, cultural differences that might impede understanding, behavioral context and signals, the intended meaning behind someone’s words, and the motivation for different types and styles of conversation.

Not to mention formatting! Another colleague shared a story with me: at the beginning of class, a student told her, “I sit in the front and use automated captions on my device. I don’t know, I might not need your services. We’ll see…”

After class, the student said:

“On second thought, I definitely need your services. Your transcript was much easier to read and follow along than the automated captions.”

Not all consumers are as good as we are at explaining what’s so great about meaning-for-meaning transcription, but they know it when they see it! And it’s not just TypeWell’s job—but that of our professional community—to demonstrate, educate, and advocate for this type of access wherever it is appropriate.

Kate Ervin is the Executive Director of TypeWell and has worked in the speech-to-text profession since 2004.