machine-readable.media

Company

Microsoft

Medical Information

Microsoft’s multi-agent diagnostician raises the stakes for clinical information tools

Strong benchmark results now face the harder test of supporting real clinical decisions safely and economically.

A peer-reviewed study reports that Microsoft’s Diagnostic Orchestrator reached 83.2% accuracy on difficult sequential cases, compared with 36.1% for generalist physicians using reference materials. Its claimed testing-cost advantage comes from simulations, the comparison was not with specialists, and Microsoft says clinical testing at a Boston hospital is still to come. If those results survive clinical evaluation, diagnostic-information products could move from passive reference support into active test selection, changing how providers buy and govern decision support.

Open Knowledge

AI answers threaten Wikipedia’s reader-to-editor pipeline while platforms buy or scrape its knowledge

Commercial answer interfaces draw on a volunteer knowledge base while reducing the visits that replenish its contributor community.

An LSE Impact essay argues that users increasingly accept answers from ChatGPT and Gemini without clicking through, reducing direct readership to Wikipedia even as AI systems continue to depend on its content. It says some companies scrape Wikipedia without authorization while Microsoft, Amazon and Meta pay for high-speed API access. If fewer readers become editors, the AI intermediary can weaken the volunteer production system that supplies its answers while shifting access and audience value toward commercial platforms.

Distribution

Copilot keeps clicked links in Edge, extending Microsoft’s control over web distribution

Mozilla’s latest tests find that browser choice still stops at the Copilot interface.

Mozilla says Windows Copilot opens web links in an Edge side panel instead of the browser a user selected in every region it tested, including the European Economic Area. Its follow-up found that European law has produced fairer treatment for some other browser choices, while Copilot routing and several Windows and Edge steering patterns remain. When an assistant controls where links open, it can redirect referral traffic and customer relationships toward its owner’s browser rather than the destinations and interfaces users chose.

AI Search

OpenAI engineer said AI users won't click source links, sharpening publishers' traffic risk

Internal remarks quoted in litigation expose the economic gap between AI citation and publisher referral traffic.

The Register reports that an OpenAI software engineer said users would not click source links however prominently they appeared, while AI answer products retain the interface built on creators' work. The column also cites a Microsoft document quoted in The New York Times plaintiffs' legal brief warning that large language models could disrupt content producers and damage their own supply chain; the court has not ruled on those claims, which OpenAI and Microsoft dispute under fair use. If citations do not produce visits, publishers gain little audience or advertising value from supplying the material that makes those answers useful.

Copyright

AI leaders knew their products could undermine the publishers they depend on

AI’s reliance on original content could become a key argument in fight for licensing revenue.

The New York Times’ legal brief quotes OpenAI’s head of ChatGPT calling AI an “existential threat” to publishers and Microsoft’s director of applied science warning of harm to AI’s own content supply, TechCrunch reports. It also cites Microsoft data showing click-through rates to the Times’ domain up to 93% lower in Copilot than in conventional Bing search. These internal warnings and figures could bolster publishers’ case for licensing payments that fund future reporting, not just compensate for past copying.