So-called 'artificial intelligence' is simply software - usually proprietary - that runs on computers.‘Artificial intelligence’1. is a marketing term;2. stands for a family of scams;3. is a social, cultural and political project. When ‘AI’ causes harms, it is usually not technology out of control, buta) software that does exactly what it is programmed to do (e.g., punishing the poor and automating the status quo)b) technology that simply does not work because it cannot work (e.g., emotion recognition). Marketing and scams serve the political agenda of the military–industrial complex, which involves presenting surveillance software as 'AI'- for the purposes of discipline, manipulation and control;- for the purposes of automation -that is, the replacement of labor with capital- to achieve a further shift of wealth from the bottom to the top. Science is a public process. Machine learning-based pseudoscience can be described as a private process due to:- proprietary systems;- training data that is not made public;- opaquely selected correlations that rise to causation;- lack of subject-matter expertise and methodological training (e.g., ML systems trained to ‘predict’ the judgement of the labelers);- obsession with quantification: theory-free and value-free ideals;- illusion of knowledge without understanding. Universities are complicit, when they act- as edtech brokers, promoting deskilling products that do not work and trapping students and researchers in proprietary infrastructures;- as corporate megaphones, propagating narratives that protect the business model of Big Tech. These narratives include myths of inevitability, the idea that there is a legalvacuum, technological exceptionalism, solutionism, anthropomorphisation of machines, dehumanisation of human beings and the use of ethics as a fig leaf and decoy. In neoliberal regimes, epistemicide was already underway.Science was being beaten to death by the ideology of merit, competitive barriers to collaboration, conformity and subjugation of researchers achieved through their prolonged job insecurity, funding for research projects on predetermined topics with predetermined outcomes, administrative evaluation of research and the 'publishor perish' culture. For once, artificial intelligence actually enables automation.The process of destroying science is automated and accelerated through - the pollution of the scientific ecosystem with fabricated data, experiments that never took place, and articles that no one wrote; - the de-skilling of researchers, who are rendered incapable of thinking and dependent on proprietary software; - the spread of probabilistic software that mimics every stage of scientific research and spreads the illusion of a science without understanding. Final question: is epistemicide, for each author of this collective crime, a desired effect or collateral damage?