The rise of citizen-to-citizen surveillance

How often are you entirely sure about what you say? 100% certain.

Probably not very often.

Maybe you even know you are wrong. You may have forgotten the exact facts, but it doesn’t really matter for the story you're telling, since the main point of what you’re communicating is still coming across.

Yes, it is essential to have a fact and science-based public discourse.

But as long as you're not an outright mythomaniac, I don’t think these minor factual errors and inconsistencies in our everyday life are a huge problem. We are not lexicons, we are storytellers with limited memory capacity.

However, the consequences of failing to verify our facts might increase very soon.

How Artificial Intelligence will create a surveillance culture

Large language models that can process both audio, text, video, and images (multi-modal LLMs) are making recording and fact-checking with the help of AI very accessible. This is likely to have a significant impact on human interactions.

Most people have been carrying audio recording devices in their pockets for almost 20 years. Practical, sometimes, when you’re doing an interview or have to remember lecture content after the fact. Nothing I’ve personally used very often.

Anyone who’s done qualitative method essays at university knows that transcribing interviews used to be one of the most painful parts of the job.

To make these audio recordings useful at scale, we need a simple way to process the information, and that is what has arrived with multi-modal LLMs. Consequently, we all now have fact-checking devices in our pockets 24/7.

If you have ChatGPT installed as an app on your phone, recording and transcribing a conversation is a simple click away.

What does this mean in practice?

From now on, every time you have a conversation or do a presentation, it can be recorded, processed by AI and fact-checked in real-time — by anyone.

And while we do have GDPR and other legislation to protect citizens online, it is currently legal to record other people, both without their awareness and their consent, as long as you are an “active” part of that conversation yourself.

Did you know?

Two types of soft surveillance, and when they appear

If you’ve spent any time on LinkedIn in the last year, people share their experience with note takers left and right — Granola, Sana, Firefly … the list is long.

I’m certain I’ve been recorded both with and without my knowledge during meetings in the last year. Those transcripts are now forever saved in databases, without my consent.

I have no idea where that information might end up, and I have no control over it.

If there is a scholarly word for this type of surveillance, I would love to know. For now, I will refer to it as “soft” surveillance. It is a different species from the surveillance by authorities, since this is primarily citizen-to-citizen.

We are likely to see two types of situations where information is recorded and processed with AI.

1. First, transparent information capture – for example, using bots and notetakers openly invited to digital meetings.
2. Second, non-transparent or secret information capture – where recordings and AI are capturing sound without our awareness.

While note-taking has been around for hundreds of years, the cost of recording and storing detailed transcripts of every single meeting is approaching zero.

How do you feel about there being an archive of everything you ever said in a professional context? Maybe even in specific private settings.

Digital footprints will soon include our offline interactions

Thinking ahead about the type of situations when this behaviour is likely to happen, I’m first of all thinking about:

• Keeping track of what was said in (1-to-1) employee-supervisor meetings, from both participants' perspectives
• Saving transcripts of information meetings with labour unions or employers to monitor potential legal or organisational concerns
• Fact-checking speakers at public talks/debates in real-time, potentially getting help to find weaknesses in their argument and raise those.
• Recording and processing transcripts of what teachers and professors say, to learn, but also to fact-check

While most of these behaviours could be neutral, or even positive, I can think of many situations where they are used more as a surveillance or control tool.

That would align with existing trends, such as cancel culture, public scrutiny, social media activism, and influencer accountability. The main difference is that these behaviours will partly move from online platforms to the offline world.

Everyday consequences of “soft” surveillance

What consequences will this have for us as citizens?

• Anyone making public statements or references will have to keep track of their sources if they get questioned on the facts.
• Everyone, especially managers, teachers and other authoritative figures, will have to be more careful with what they say, to ensure it is not possible to interpret incorrectly.
• This might affect spontaneous, open and creative conversations, even in private, since our words might be saved or fact-checked without our knowledge.
• It might reduce the amount of controversial but constructive and legitimate academic or political discussions, since it might be easy to challenge and criticise alternative views.
• We might all over-optimise our communication toward safe and "AI-friendly" content without nuance, avoiding new ideas.

Societal consequences of “soft” surveillance

I don’t think a society where we surveil each other, for any reason, is a healthy society.

While you could argue that fact-checked public discourse could help reduce misinformation, conspiracies, and propaganda, I think a massive part of being a human thinker is about testing ideas that might not hold up when scrutinised.

Being wrong and challenged in friendly settings is what moves our thinking forward.

We also encounter many situations where the truth or facts are not black and white. The majority of interesting conversations take place in the grey zone.

As someone who used to be a manager, I would have dreaded interacting with employees if I knew that every challenging conversation we had would likely be recorded and potentially reported, even in one-on-one settings.

How to reduce the risk of “soft” surveillance hurting you personally

Personally, I’m starting to prepare for this in a few ways:

1. I assume it is all recorded already. And I try to keep a good track of my sources and claims.
2. I rarely record anything, and I’m always transparent if I am recording something (and you should be too)
3. I’ve actively decided to take the risk of being criticised and fact-checked, and I will continue to publish and share my thoughts in public.
4. I regularly clean up my digital track record and limit my digital presence
5. I try actively to work on creating environments where we are open-minded and listen with good intentions, instead of being critical and finding faults