AI Does Not Need to Be Fascist to Serve Fascism

Calling AI fascist gives the technology far too much agency. AI has no political ideology. Databases are not authoritarians, cameras do not believe in dictatorship, and models do not wake up yearning for the suppression of democracy. But technologies do not need political convictions to become extraordinarily useful to political projects.

AI can make surveillance cheaper, classification faster, propaganda easier to produce and personalize, administrative decisions easier to automate, populations easier to sort and enormous amounts of information easier to process. Democratic governments can use those capabilities. Companies can use them. Authoritarian governments can use them too, with considerably fewer restraints. The danger does not require a machine that wants fascism. It requires institutions with authoritarian ambitions gaining access to infrastructure that makes previously expensive forms of control cheap enough to operate at scale.

So “Is AI fascist?” is not a particularly useful question. Ask which forms of power it makes cheaper, faster and easier to scale. Ask who has access to those capabilities, who can challenge their use, which institutions are supposed to constrain them and what happens when those institutions themselves begin to erode. Technology does not need to share the ideology of the people using it. A machine can be politically indifferent and still become extremely useful to people who are not.