The Manifesto on Algorithmic Sabotage is guided by three core principles:
Hmm, the term "algorithmic sabotage" suggests a critical or even adversarial stance towards algorithms, likely in the context of social media, recommendation engines, AI, or digital platforms. The user might be interested in digital resistance, tech criticism, or counter-strategies to algorithmic control. They could be an activist, a writer, a scholar in media studies, or someone frustrated with algorithmic systems. Their deep need is probably not just information, but a compelling, structured argument that can serve as a call to action or a framework for thinking about resistance. manifesto on algorithmic sabotage
The algorithm wants to predict you. It feeds on your consistency. Sabotage begins by being unpredictable. Click on what you "hate." Ignore what you "love." By poisoning your own data profile, you become a ghost in their marketing machine. If they cannot categorize you, they cannot own you. 2. Practice Generative Friction The Manifesto on Algorithmic Sabotage is guided by
What happens if we succeed? If we poison the data enough, the models will enter a state of . They will begin to feed on their own previously sabotaged outputs, creating a fractal spiral of nonsense. Their deep need is probably not just information,
More importantly, the act of sabotage changes the saboteur. It transforms passive subjects into active agents. It replaces helplessness with intention. It builds the psychological infrastructure for collective action. The sabotage itself is training for resistance.
Further Reading: "The Age of Surveillance Capitalism" by Shoshana Zuboff, "Weapons of Math Destruction" by Cathy O'Neil, "Automating Inequality" by Virginia Eubanks, "The Stack" by Benjamin H. Bratton
Algorithmic sabotage is the intentional degradation of a machine learning system’s performance, reliability, or truth-output. It includes but is not limited to: