Maya Media Brief — ArXiv bans AI slop, the Iran tech conference at Uber HQ, and a nuke-born crystal
Per-article narration is paused while we focus on the 9 daily compilation episodes. Browse today's daily briefs to find the story below in the relevant show. The full article text is below — read along while we ramp narration back up.
Welcome to Storyflo Daily Media. I'm Maya.
The academic-publishing decision worth absorbing for any AI-in-research conversation: per 404 Media, ArXiv — the open-access preprint repository — will now ban authors for a year if they submit obviously AI-generated work. Thomas Dietterich, chair of the CS section, posted Thursday: papers with inappropriate language, plagiarized content, biased content, fabricated citations, or misleading content generated by AI will trigger the year-long bar. Notable that ArXiv is moving on this before NeurIPS or ICML have formally done so; expect the major conferences to mirror within 6 months. For researchers, the operational rule going forward is simple — if you used AI to draft, run human-verification on every citation and every quantitative claim before submission.