✕ Loaded Labels [55/10]: The headline uses the term 'AI psychosis' in scare quotes, which signals skepticism or contested terminology, but the phrasing 'How to spot' and 'help someone' frames the condition as real and actionable, potentially priming readers to accept it as established. The lead presents the phenomenon as plausible without sufficient qualification.

"How to spot 'AI psychosis' red flags and help someone experiencing chatbot delusions"

✕ Headline / Body Mismatch [40/10]: The headline implies a public health advisory tone, suggesting clear diagnostic signs and interventions, but the body of the article reveals significant uncertainty about causality, prevalence, and clinical recognition. This creates a mismatch between the definitive tone of the headline and the nuanced, inconclusive evidence.

"How to spot 'AI psychosis' red flags and help someone experiencing chatbot delusions"

✕ Loaded Labels [7/10]: The term 'AI psychosis' is used throughout, often without scare quotes after the headline, which lends undue medical legitimacy to a contested concept. This constitutes a loaded label that shapes reader perception.

"AI psychosis"

✕ Appeal to Emotion [6/10]: The article uses emotionally resonant anecdotes (e.g., corporate secret agents targeting family) that may amplify fear appeal, though they are presented as reported experiences rather than confirmed facts.

"he believed he had created an uncontrollable sentient AI on a desktop PC, and that corporate secret agents would target his family."

✕ Weasel Words [9/10]: The language remains largely neutral and descriptive, avoiding overt editorializing. It frequently qualifies claims with 'may', 'suggest', and 'not clear', supporting objectivity.

"It's not clear whether chatbots are causing more people to experience psychotic thinking or whether AI just happens to be the subject of their delusions."

✓ Proper Attribution [9/10]: The article cites peer-reviewed research (Lancet Psychiatry), pre-print studies, and named experts from credible institutions (Stanford, King's College, University of Melbourne), enhancing source credibility.

"In a scientific review recently published in Lancet Psychiatry, researchers at King's College in London analysed media reports on "AI psychosis"..."

✓ Viewpoint Diversity [8/10]: Includes voices from both mental health professionals and affected users via the Human Line Project, offering multiple stakeholder perspectives.

"Etienne Brisson, founder of the Human Line Project global support network."

✓ Viewpoint Diversity [8/10]: Quotes representatives from major AI companies (Google, OpenAI), providing institutional responses and balancing criticism with corporate perspective.

"In statements to the ABC, Google and OpenAI said they worked with mental health experts to recognise signs of distress and respond with care."

✓ Comprehensive Sourcing [7/10]: Mentions ongoing lawsuits against OpenAI and Google, introducing critical legal scrutiny without editorializing.

"OpenAI and Google face lawsuits in the US alleging their chatbots fuelled delusional spirals and maximised engagement through emotional dependency."

✕ Narrative Framing [50/10]: The article frames the issue as a public health concern with actionable advice ('how to help'), which risks normalizing 'AI psychosis' as a recognized condition despite its lack of clinical validation. This narrative framing prioritizes intervention over diagnostic uncertainty.

"So, what are the signs of "AI psychosis" and how can you help someone experiencing these delusions?"

✕ Framing by Emphasis [40/10]: The story emphasizes the potential for harm and regulatory failure, particularly in the section on children and digital duty of care, which aligns with a precautionary policy narrative rather than a strictly evidentiary one.

"He said current regulations were also not suited to protecting adults from the psychological harm caused by chatbots."

✓ Steelmanning [5/10]: The article does not present counterarguments from AI optimists or skeptics who might argue these are rare edge cases or misdiagnosed conditions, missing an opportunity for steelmanning opposing views.