Emily Bender and Nanna Inie published a guide on discussing AI technology without using language that falsely suggests these systems have human-like qualities — a documented problem in AI research where over half of published studies assume consciousness-like traits before actually testing for them. The piece addresses a real structural issue: anthropomorphic framing (calling systems 'intelligent,' 'understanding,' 'believing') obscures what these models actually do — pattern-match and generate text based on training data. The guide helps readers separate the marketing language from the technical reality, which matters because clearer terminology makes it harder for companies to overclaim capabilities or for policymakers to misunderstand what regulation is needed.