A fellow AI professional recently told me, with some conviction, that using an AI note-taker without disclosure is illegal. My own understanding was different. If you are using it for your own purposes, I thought, that is broadly fine. If you are sharing the transcript with others, that is where it gets complicated.
So I looked into it more carefully. What I found was more nuanced than either of us had appreciated.
I use these tools myself and think they are genuinely valuable. But I have been asked to turn mine off once, in a conversation where personal information was being shared. That instinct told me something worth paying attention to.
The understanding most people hold, that personal use is fine and sharing is the issue, made sense in a world where recording meant a file that stayed on your device. AI note-takers broke that model without most people noticing.
What the tools actually do
When one joins a call, it simultaneously records, transcribes, and uploads conversational data to a third-party cloud system in a single default action. The processing is the tool. And most participants have no idea that any of it is occurring.
The legal implication is where it gets interesting. Automatic upload to a third-party system triggers transparency obligations under UK GDPR whenever personal data is being processed.[1][2] A reference to an employee’s performance, a client’s situation, a colleague’s name, all of it qualifies. I am not a lawyer and specific situations warrant proper advice, but the direction is clear enough to matter.
The tools themselves can’t agree on what disclosure looks like. Some appear as a named participant, visible and impossible to miss. Others drop a line into the chat, technically disclosed, practically ignorable. Others, including hardware devices like Plaud that capture physical meetings and phone calls, leave the other person with no indication at all.
In agentic systems the conversation doesn’t just get transcribed. It gets analysed, summarised, and fed into automated workflows without any human decision point in between.
The law was built for a world where recording, processing, and distributing were discrete intentional acts. That world no longer exists.
A simple professional standard
These tools are worth using. They deserve a simple professional standard to match: disclose that you are using one, every time, before the conversation starts. Not because the law currently requires it in every case. Because the people you are meeting with deserve to know.
She was wrong about the legal position. So, it turns out, was I.
If you are working out how your organisation should use AI note-takers and similar tools responsibly, I would be happy to help. You can read more about our services or get in touch.
This article was first shared as a post on LinkedIn.
Sources
- Information Commissioner’s Office, Right to be informed.
- Information Commissioner’s Office, How do we ensure transparency in AI?