Transcription etiquette: 4 rules for AI in meetings

Unknown stinky face present

Let’s talk about transcription etiquette.

Transcription tools now come in all shapes and uses. From a quick summary to a full transcript of a meeting, from brain-dumping to catching up on a meeting you missed. There are countless ways to use AI transcription tools, and I use them regularly myself.

Before we get to all the benefits and uses, there are a few things you need to know.

Transcription etiquette: four basic rules

1. Always ask for consent before you start a transcription tool. Everyone in the meeting has to agree to their data being processed in an AI tool.

2. Explain what the tool is being used for. Preferably twice: before you ask for consent, and again once everyone has agreed and you start the transcript. That way it is clearly recorded in the transcript as well.

3. Know what happens to the data. You need to be able to explain where the data goes, whether it is used to train models and how long it is kept.

4. Give people real room to say no. Without social pressure and without it straining the relationship.

If you do not know the answers to points 2 and 3 yourself, my advice is simple: do not use the tool.

When you are a participant and are asked to consent to a transcription tool, you can refer to these points.

Starting transcription automatically as soon as a meeting begins can also conflict with European privacy law, the GDPR, which applies in every EU member state. If a tool starts automatically, a participant may no longer be able to object before their data is processed. In many tools this even applies to messages in the chat.

So using transcription tools is not only about efficiency, but also about trust in a conversation. People need to feel free to speak without wondering where their words will end up later.

Back to the stinky face

During an online meeting, the organiser joined the call early and was talking to his dogs: “Who’s a lovely stinky face? Come here you stinky face!”

The transcription tool started automatically as soon as the meeting began and afterwards, without any human check, sent a summary to all participants.

The first sentence of that summary:

“Unknown stinky face present.”

Funny? Absolutely.

But misunderstandings like this creep in surprisingly easily, especially in automatically generated summaries.

You might, for example, talk to a client about a project for twenty minutes. Halfway through, the client says: “This isn’t quite what we had in mind.” Part of that conversation can still end up in the summary as a fact or a decision. Sarcasm? Hard to recognise. Humour? Often the same problem.

And a conversation about someone who is not present? If that person was invited to the meeting, they may well be able to read the summary.

In short: always read through the transcript and the summary.

AI and language: sometimes comical, sometimes confusing

I regularly sit laughing behind my screen when a transcription tool tries to make sense of English jargon in a Dutch-language meeting.

GenAI, for example, turns into “Jenna” or “Jenn & I”.

During a training session I demonstrated a transcription tool (and yes, I neatly followed the four steps above) and told the group about Unknown stinky face present.

The tool translated this beautifully in the Dutch transcript as “Stinkend Feest”, which means “Smelly Party”.

Where transcription tools are genuinely useful

Fortunately, this does not mean transcription tools are a problem. Quite the opposite. They can be hugely valuable when you use them consciously.

Media

• Turning spoken content into written content
• From video or podcast to blog
• Writing up interviews for publication
• Generating subtitles for videos

Live events

• Capturing participants’ insights
• Documenting Q&A sessions
• Analysing discussions
• Improving training material after a session

Meetings

• Generating minutes
• Recording action points
• Documenting decisions
• Being able to find conversations again later

Knowledge management

• Building internal knowledge bases
• Archiving important conversations
• Developing onboarding material

My personal favourite

At the end of the day, or in between, I often use transcription for a quick idea dump. I simply say out loud everything that is in my head. Then I let the tool organise it and turn it into an overview or action list.

What do transcription tools mainly give me?

The biggest benefit for me is simple: I can listen more attentively. I still take my own notes during meetings, because writing along helps your brain process information better. I can also add nuance straight away to what would otherwise be a fairly flat transcript. So transcription tools are not a replacement for thinking or listening, but they are powerful support.

And the nice thing is: support for your minutes is only one click and mutual consent away.

When you stick to good transcription etiquette:

• the possibilities are enormous
• it can genuinely save you a lot of time
• “Unknown stinky face present” will never happen to you

Getting started

Want to use AI tools responsibly in your organisation? We can help with AI consulting for organisations and practical AI training courses.