Nightshade
A friend works with extremely sensitive data. Their employer issues them a phone, which they use indiscriminately. Their employer is subject to FOIA requests. Their tech usage means my conversations with them, my activities with them after hours, their associations with technology I control, is all a matter of public record.
This person's role is technical-adjacent; they know about behavioural analytics. They've certainly gotten static about it from me. So, why not get aggressive about blocking, why not limit the information available to this device to work-relevant data only?
Because I need to fit in.
They need to generate the same sort of data that everyone else generates, otherwise the people who watch them watch harder. Harder watching makes watchers more likely to invent things to find, and that means trouble.
This is a chilling fact about Ubiquitous Benevolent Surveillance. Here as well, science fiction is decades ahead. Philip K. Dick's 1977 novel A Scanner Darkly muses on exactly what life under what was then considered an impossible all-seeing intelligence might be like.
As an aside, a word for "all-seeing" is pantoscopic. Jeremy Bentham borrowed this notion for Panopticon – his arguably terrifying idea for a prison in which all the cells are arranged in a circle around a central tower. Bentham was writing during the Enlightenment, when the application of "scientific" ideas to social phenomena was vogue – in this case the idea that observation alters behaviour of complex systems, here applied to the problem of corrections. The term "pantoscopic" mostly gets invoked in technical discussions of lens groups for photography, a fact less shocking than it probably should be.
Anyway, part of A Scanner Darkly's premise is that undercover agents are unknown to even the organization that they work for. The novel follows a character's descent into drug-fueled paranoia as a result of his work. He must break the law in order to maintain his cover, and the ubiquitous anonymous surveillance of his crimes drives his paranoia.
This character's need to fit into a druggy subculture to do his job becomes the very thing that presents the existential threat. That anyone in his social group could also be an undercover agent – could be collecting data about him for the purposes of prosecution – drives him into a paranoid psychotic break.
Philip K. Dick understood the chilling effect of social media before the people who made this machine had even been born.
So, like the main character of A Scanner Darkly, my sekrit squirrel friend uses their work phone like their colleagues do, including activities that a reasonable, privacy-conscious person would not want surveilled. It means they must be constantly vigilant about where they go, what information they consume and transmit, and what they do with this device. Sure, this is arguably a consequence of the choice of occupation – consider members of the military who have revealed "black sites" through their fitness data. The technology imposes a duty of care that requires significant consideration, is the thing.
Accepting work that requires secrecy means that secrecy extends to one's personal life to an extent that technology’s use must be considered thoughtfully by every individual who does so. This same surveillance, however, is willfully participated in by basically everyone, with the same chilling effect. The difference is in consequence.
So the question about how much data to share is a complex one. Even for us regular people unburdened by tradecraft. As with all security considerations, zero is the instinct. Complete removal from and refusal to participate in profiling and data collection is often set as the goal. This has material impact on one's quality of life and ability to fit in with one’s social group – imagine the challenge of being a member of Gen-Z without access to Snapchat or TikTok – the modern equivalent of that weird kid in your middle-school class who wasn't allowed to watch television.
Non-participation is difficult, silent, and easy to contravene. Instead of direct profiling, advertisers use shadow profiles to collect data about users who actively interdict tracking technologies. When your stock android-using friends join your home wifi network, their profile data is associated with yours and surprise! your profile gets updated with their data. Most shadow-profile activity gets misattributed to always-on speech-to-text technologies. No one is doing that. Association metadata is far more available, easier to collect, and faster to work on.
The better solution than non-participation is to jam these efforts. Random data in any dataset makes it inefficient, difficult to parse, and valueless if done at scale. Data poisioning usually gets discussed in the context of machine learning, but it's been a thorn in the side of even basic trend analysis in podunk analytics tools.
So, poison your data.
Analytics data mostly exists in JavaScript. That's client-side. You're the client. It's your processor cycles that get used to generate and transmit data about your activities. Why not exercise some agency.
Browser extensions like Noiszy don't work to halt data collection. They stuff it full of the exact bullshit that breaks datasets. Alternatively, go here and press a button. Go do it on every computer you have access to. Make some fucking entropy.
If nothing else, eventually your hardware won't be involved in your surveillance any longer.
There are server-side tracking tools, sure. The industry is slowly moving toward them, but they're not as slutty with your information, so adoption has been slow. Why not help that along.
Force us to innovate. Make us work for our trespass.