A brief insanity

So I have a theory that we are very soon going to end up in a state where we all simultaneously have complete privacy and zero privacy, and therefore privacy will cease to matter much at all.

First: I am not talking about safety. Safety and privacy are related concepts but not the same. If someone with nefarious intentions knows your location, that is a privacy breach that creates a safety problem.

I am talking about privacy in relation to the current fear we have that some private aspect of our lives will become public. I’m not sure what happens to that fear as AI becomes better and more widespread, and I think it’s possible it goes away. Say I’m a politician and someone filmed me doing something my voters would disapprove of. If that video is released in 2015, it’s a huge problem for me. If it’s released in 2028, I’ll probably just say it’s AI. Say that I’m cheating on my spouse, and somebody sends them a video of me on a date with another person. At any other time in history, that’s almost certainly proof. In 2028, I’ll probably just say it’s AI.

If it becomes impossible to discern real video from AI-generated video, I think the current privacy concern we have Meta glasses will disappear. (I suppose it’s possible people go the other way, and try to never allow their image on the internet so that no-one can ever use them in AI-generated content. But I suspect that will be too hard to do in reality, and we’re probably too far gone in terms of posting our image on the internet anyway).

This brings me to my second theory: there is probably a lot of useful information we could learn about ourselves if we had a visual log of everything we do. I’m talking health-wise. I have a belief that the majority of the way our body operates is directly related to our actions and external environment. If I eat a certain thing, do a certain exercise, or use a certain product, it can affect my body in both positive and negative ways.

It is currently almost impossible to track those effects because pretty much no-one (except maybe Bryan Johnson) has a consistent record of the external inputs to their body, and the internal outputs. We’re getting better at tracking the internal outputs via wearables that measure heartrate, sleep quality, etc. There are lots of interesting projects on the internet (and at wearables companies) of people trying to track correlations between these things. But we’re missing the causation piece. If my heartrate spikes randomly, that data alone says nothing about what might be happening - it could be because I’m exercising or in a stressful meeting or I drank alcohol recently. If I have bad sleep quality, my Apple watch has no idea what time I ate my last meal or whether I had bright lights on before bed or whether I drank coffee a little late in the afternoon.

My hypothesis is that if it were possible to comprehensively track all the time series data of a person’s actions/external environment, plus measure their body’s internal state, you could train a machine learning model or neural network to predict personal outcomes based on actions/environment. God tier personalised medicine. And in aggregate, we could probably learn a huge amount about the human body. The hard part about testing this is that no normal person is actually tracking that level of data.

In a world where everyone is wearing Meta glasses or some form of wearable, tracking those external circumstances becomes much easier. I wanted to test this theory on myself, so I bought a police body cam off Amazon (I didn’t know you could do that either) and wore it around for a week. I also bought a Samsung watch to wear. I had the camera take a picture every few seconds. At the end of each day I would upload the pictures to my laptop. I’d run a few scripts that removed any images that hadn’t significantly changed versus the previous image, and then classified the remaining images into categories of actions and tagged them.

This was an interesting thought experiment but I pretty quickly realised that even I, who definitely sits on the extreme end of wanting to optimise my health and collect data on myself, found it annoying to have to wear a camera around everywhere. Meta glasses might have been less annoying but society is for sure not at the point of accepting them (at least not where I live). Plus I have prescription glasses I need to wear instead. After about a week my focus shifted to other priorities, so I never collected enough data to really answer my question.

The concept is still something I’m wondering about, but I also think most humans would find it tiring to track so much data about themselves in the name of optimisation, even if it were seamless. There is something to be said for an unoptimised life being the root of creativity and diverse experience. But if the Meta glasses or similar do become socially acceptable and not annoying to wear, I’ll probably revisit this experiment regardless.

PLEASE BE AWARE: I did google the law on this beforehand and it was legal in public places in Australia as long as I didn’t record audio of conversations without people’s consent (obviously don’t take this as legal advice I’m not a lawyer). Also note that I was working from home during this time and very rarely left my house, so almost all of the images captured were just me doing things in my home. When I did leave the house, the camera was angled slightly toward the ground so as to capture what I was doing without capturing other people’s faces. All images were deleted at the end of the project.