Attempts to Ban Face Recognition and ML Will (and Should) Fail

You can’t ban facial recognition technology as this BuzzFeed screampiece suggests. You can’t ban building a model and querying the model, neural networks, or recording digital video. That genie has left the bottle and is on a plane to Singapore sipping mimosas.

You can ban misuse of the result of those models, but we already have laws for that.

Sure, facial recognition is going to be used by authoritarian regimes to oppress and enslave us. Last I checked, so were tanks, nukes, taxes, water rights, and every single piece of technology humans have ever created.

Stop thinking that the problem is technology, and start thinking the problem is naked apes in silk suits pretending they are leading while they are just grunting in favor of their own tribal power.

So why should we think face recognition or machine learning tech isn’t terrifying? One reason: there are thousands and thousands of GitHub repos where normal programmers are hacking and reverse engineering the future. Because unlike tanks and artillery, ML can be used against the ruling class just as it can be used against us.

Here’s an example: let’s say I want to prove that a Politicians or CEO is lying. I know, a real edge case.

The way it works now is we wait 24 hours for some intrepid reporter at huffpost or wherever to dig up the video, and… we have already moved on as a society (and the liar has gotten away with it). All this depends too much on people who care knowing what happened in the distant past, and people who care going to a website and caring enough to care.

In the computing world, we would say our monitoring and alerting system for truth is not alerting fast enough. Thanks journalism, but turns out you are waaaaaay too slow for the 21st.

But what if everything a professional liar ever said was indexed and put into a model of them, where a system that understands their words would instantly pull up video of their previous lies alongside their current bullshitting. ML could allow us an experience of multiplexed video of those people through time.

In other words the next time a liar like Zuckerberg lies in front of Congress, we can display realtime his previous lies which expose his new ones. Like this 2009 video of him saying Facebook will never sell your data.

Imagine having to give testimony when your previous lies are blasted in your face. Go ahead, argue with yourself for a while. Imagine a reporter holding up their phone and asking questions in real-time based on what an ML analysis of the liar’s words were.

I would guess that liars would hate that. Good, let’s drive them into the darkness.

Quit pretending that this is about appealing to reason, or protecting our imaginary rights. This is historical, epic war, and to win against the oligarchy we are going to need to have our own suite of metaphorical weapons. ML is going to give us the tools to change the game, and secure a future based on supporting the health and happiness of individual humans, Singletons, which I would argue isn’t socialism at all.

It may take fifty years, or a hundred, but society will evolve around this tech as it always does. But if try to ban ML, you’re taking away our best technique of making this change equitable.

Because real equality needs real data, and tools and people to understand it.

Render by Mike Winkelmann

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.