"who owns the machine?"
‘who owns the machine?’ is a conversation between peter michael witt and the ai system claude (anthropic) about a gap in the current discourse: the public debate on artificial intelligence revolves almost exclusively around threat scenarios – mass unemployment, concentration of power, the erosion of democracy – without ever asking an obvious question: if the machine is so powerful, why doesn’t it belong to all of us? the interview explores this question across twelve question-and-answer
segments: from günther anders’ concept of the obsolescence of humankind and mark fisher’s ‘capitalist realism’, through salvador allende’s forgotten cybersyn experiment, to the epistemic limits of hayek’s information argument.
the format has a performative dimension: a human being questions an ai about the socialisation of ai – on an infrastructure owned by a private american company. the first question is therefore: ‘can i even trust your answer?’ this
contradiction is not resolved, but forms a conceptual core of the conversation.
the text emerged from a larger research project. a four-part essay on the same subject – featuring a detailed reconstruction of cybersyn, a model sketch of a collectively controlled ai economy, and what are surely necessary legislative initiatives from the eu – is intended for a separate publication. the interview stands as an independent text, distinct from this.
the text was published in september 2026 in „merkur – deutsche zeitschrift für europäisches denken“
