It’s time to write a bit more about privacy and surveillance. This is without a doubt the topic that people have found the most interesting and concerning – in my daily conversations, in my online discussions on LinkedIn with other academics, and also, in the news.
Every day it seems, a new story about Meta, or other AI glasses being used for spying, voyeurism, or generally being creepy turns up.
The best summary I’ve found of the issues is on the Electronic Frontier Foundation’s website: “Think Twice Before Buying or Using Meta’s Ray-Bans“. For me, the game is slightly different – I am explicitly not recording others. My conversations however, are uploaded to Meta. This is a problem I have sought to make my peace with, especially since my conversations are probing the AI capabilities, rather than asking about deep, personal questions.
But people in the supermarket don’t know that. In my head, I hope I come across as a friendly, offbeat, slightly eccentric type who fits perfectly into the role of ‘academic researching wearable technology’. I fear that my mental image of myself is likely not how others perceive me. I could well be seen as a smartglasses creep – and I don’t want that.
“Who is that guy muttering to himself in the milk aisle about why people buy Jersey milk instead of regular? Just a creepy AI-glasses guy”.
The thing that sticks out to me is, why are AI glasses with cameras so much more controversial than a) having a phone, which you can use to record people with and b) spy cameras, which have been around forever.
I have a theory.
Firstly, the positioning. The camera sits at eye level. A phone is held out to record. You can see where it is, you can potentially step out of the frame. You can read the intention of the recorder more clearly. The glasses are in the same space as the gaze – they follow wherever you look. We are sensitive to being watched, or being observed. The idea of a person both watching and recording you from eye level is in my opinion, a unique threat that speaks to something very ancient in our minds.
There is also an asymmetry at work – there’s a reason that in those online videos you see of altercations and road rage incidents, whenever one person is recording, the other person will comment on it and even start recording themselves. The glasses maintain this asymmetry, but it is not possible for the other person to ‘match it’.
Even the LED light that comes on when recording – I have not actually used this, but I can imagine that even though it alerts those around to the fact that recording is happening, it modifies the social interactions happening around the person.
So how about spy cameras? In order to address this, I’ve completely ruined my Google algorithm, which I have spent years carefully curating. It now thinks I want to buy spy gear. But it was worth it, as it confirmed my suspicion that there are far more sinister tools available for recording secretly. Just like we have been seeing in the movies for the past 30 years. Yet these do not seem to get attention like the glasses are getting attention. Why?
My theory:
1) These are niche products, and I’m not sure who they are for. I would like to think security teams etc. but maybe I am being too optimistic.
2) These do not have plausible deniability. If you are caught with something like that, you would not be able to explain it away. If you are wearing socially acceptable new AI tech which just happens to record, and you ‘forgot’ you were recording, there is more plausible deniability.
3) These are secretive, so people don’t know. I am now feeling much more paranoid about being spied on in public.
The social issues with observation, privacy, and surveillance though, go further than just voyeuristic observation without consent.
It is almost obligatory at this point to mention Michel Foucault.
The panopticon, the famous prison design that Foucault invoked to explain surveillance, was built on the fact that the prisoners could always, in theory, be observed from a central watch tower. They would have no way of knowing if the guards were there or not. The possibility of surveillance itself was a form of power and control (a common mistake is thinking that it’s constant surveillance that matters, it’s not – it’s the uncertainty of when you are being surveilled that is important).
The more I look into this, the more people I speak to, and the more I read online, it occurs to me that the AI assistant is simply a stowaway in a bigger converstaion about surveillance and privacy. The issue isn’t AI glasses or smart glasses – it’s camera glasses.
I wonder, if we are consistenly observed, or under the threat of observation from others, through wearable cameras – do we feel the weight of being surveilled?
What would Foucault say about AI glasses with cameras?
If I guess, I would say that it is not that the cameras, nor the panopticon is ‘bad’.
I think he might consider these points:
1. In a panopticon, the observation and uncertainty of observation is centralised and run by an authority. With AI Glasses, it is distributed among individuals.
If we are all wearing camera-enabled glasses, then the gaze is distributed among all of us. Each wearer is their own panopticon. Surveillance, or the threat of surveillance, is pushed from the top down (CCTV cameras belonging to institutions) to street level (the individual’s eye line).
2. These devices produce knowledge effects, which produce power effects
The observation and surveillance capabilities of AI glasses are not neutral. Especially, if they can contain facial recognition technologies. They will produce information, automatically, which becomes data, which can produce leverage, discipline, or threat.
3. Productive Power and Biopower
We are not forced to buy these. We want to do these things. We want to buy new technologies that turn us into surveillance-machines. We buy them, because they disguise surveillance and datafication as pleasure, fun, and amusement. At the same time, we cede control over data monitoring functions of our body (for example, blood pressure in smartwatches, our vision in smart glasses) to technology and technology corporations, who become custodians of the data (and who often misuse it).
4. Technologies of the self and confessions
Foucault talked later in his writing about ‘technologies of the self’:
“Technologies of the self, which permit individuals to effect by their own means or with the help of others a certain number of operations on their own bodies and souls, thoughts, conduct, and way of being, so as to transform I themselves in order to attain a certain state of happiness, purity, wisdom, perfection, or immortality.”
The link here really writes itself.
- We try to achieve greater happiness or social satisfaction by posting, photographing, documenting our lives for others’ approval.
- We try to transform our thought and make it more efficient and precise.
- We try to reduce friction through technology by becoming ‘perfect’ through the lens of efficiency and productivity.
- We try to reduce uncertainty (I can’t tell you how many times I’ve checked the time or the weather on these things).
- We might not be chasing ‘happiness’ directly, but no one spends £400+ on these in the hope of achieving a worse quality of life.
Coming back to the idea of standing in the milk aisle and talking to myself, the main thing that concerned me wasn’t getting the right information from my glasses, it was how I’d be seen. I have installed a panopticon-style device on my face, paid for it, and now as I try to use it, I am considering policing my own public behaviour (not using the AI) because I am aware that I am being observed; or maybe I’m not being observed. The panopticon in miniature.
So I end with asking – what exactly are we buying, or buying into?
Leave a Reply