Welcome to Tech Beat, your daily look at the stories shaping our digital world.
ArXiv is drawing a hard line on AI-generated slop in academic research. The preprint platform will now ban researchers for a full year if their papers contain hallucinated citations or unreviewed AI output left in plain sight. It is a notable signal that the scientific community is losing patience with researchers who treat language models as a substitute for actual scholarship.
Meanwhile, the relationship between Silicon Valley and government is getting harder to ignore. Palantir, the data analytics firm with deep ties to US intelligence, has quietly hired more than thirty senior UK government officials. Critics argue that kind of talent pipeline blurs the line between public service and private interest in ways that deserve far more scrutiny than they typically receive.
And in a genuinely strange corner of artificial intelligence, French researchers have used a language model trained on the works of Molière to produce an entirely new theatrical comedy. The play premiered at the Versailles Opera more than three hundred fifty years after the playwright's death. Whether that counts as art or a very elaborate tribute act is a question audiences are now being asked to answer themselves.
Keep surfing. Tech Beat out.