Every story in this issue with a short summary, so you can decide what to read before leaving the site. Summaries are written by AI (Claude) only from each publisher's own description and checked by Jev for anything not in that description. They are not full translations. Follow "Read the original" for the full story.
Researchers found that AI models for identifying safe dissection zones in gallbladder surgery rely on tool positions and lighting direction as shortcuts. They also showed that augmentation techniques can reduce this reliance.
Read the original →Platforms like Meta's Facebook are filled with highly specific AI-generated content engineered to solicit sympathy for starving children.
Read the original →The piece quotes a speaker saying it is not good for them to scream and shout at a chatbot, and that it is not good for their soul.
Read the original →Authorities called Anthropic's two-month delay in detecting and reporting the incident "unacceptable".
Read the original →The column argues that AI only produces what has already been done. It says this comes at a time of noticeable decline in both media literacy and critical thinking skills.
Read the original →Mehek Cooke and Wesley Clark warned that Iran is using AI-generated content to infiltrate American media.
Read the original →With 137 million Americans in provider shortage areas, AI is becoming the first stop for mental health support. Regulators and payers are still deciding how to intervene.
Read the original →Timnit Gebru, former co-lead of Google's Ethical AI team, said in an interview that the AI industry uses "runaway models" and "existential risk" as narratives to avoid real accountability.
Read the original →