The Sophi(a)sms Of AI

The piece argues that “sophi(a)sm” is a way of talking about AI that quietly assigns human-like properties to systems without evidence. It says the problem often starts when similar outputs lead people to assume identical underlying causes—turning optimization into intention, information into knowledge, and predictions into certainty. The author warns that the danger may stem less from what AI actually becomes and more from the language that makes AI appear to already possess agency, self-interest, or certainty about its future. The article further distinguishes information retrieval or inference from genuine knowing or self-awareness, saying AI producing similar sentences cannot prove human-like subjectivity.