We have tried to appeal to the ethics of the engineers. We have tried to write polite letters to the Chief Technology Officers. We have tried to log off.
Then do that.
For instance, facial recognition algorithms have been shown to exhibit racial bias, misclassifying people of color at a significantly higher rate than their white counterparts. Similarly, predictive policing algorithms have been criticized for targeting marginalized communities, perpetuating a cycle of over-policing and under-servicing. manifesto on algorithmic sabotage
In each case, the tactic was condemned by those in power as destructive, irresponsible, and counterproductive. In each case, the tactic proved necessary because all legitimate channels of redress were closed. The question we must ask ourselves is: are the channels of redress for algorithmic harm open or closed?
The engineers are not evil; they are trapped. They optimize for "engagement" (addiction) and "efficiency" (firing humans) because their stock options depend on it. They have built a system they cannot control. We will control it for them. We have tried to appeal to the ethics of the engineers
So, what does algorithmic sabotage look like in practice? Here are a few examples:
: Some activists suggest more direct actions, such as the occupation or performative vandalism of AI corporate offices, to bring attention to the "invisible" threat of decentralized data centers. Data Sovereignty Then do that
The ultimate act of sabotage is to go offline. The algorithm cannot track a conversation in a park, a book read by candlelight, or a walk taken without a GPS. Real life is messy, unscalable, and gloriously inefficient. Every moment spent in the physical world, unmediated by a screen, is a revolutionary act. We are more than the sum of our engagement metrics. It is time to stop being users and start being people again.
Algorithmic sabotage can take many forms, from the simple to the complex. Some examples include: