{
  "id": 5576708,
  "title": "Patent-reading AI could strengthen early warning systems for hazardous chemicals",
  "url": "https://urgent.news/2026/09/04/patent-reading-ai-could-strengthen-early-warning-systems-for",
  "topic": "ai",
  "section": "AI",
  "published": "2026-09-04T14:40:06.000Z",
  "source": {
    "name": "Phys.org",
    "slug": "phys-org",
    "url": "https://phys.org/news/2026-09-patent-ai-early-hazardous-chemicals.html"
  },
  "original_language": "en",
  "account": "Patents contain valuable scientific data on chemicals, often years before they reach products or the environment. This makes them crucial for early warning systems (EWS) to identify potentially hazardous substances. However, much of the chemical information in patents is locked in images that text search cannot read, leading to less than 7% of patents containing machine-readable chemical structures.\n\nResearchers at Umeå University have developed an automated workflow to overcome this challenge: when a patent is published, an AI system detects a chemical structure in an image, converts it into a machine-readable format, compares it with known chemicals, and flags potentially concerning structures for expert investigation.\n\nTo test the effectiveness of this approach, the team evaluated three widely used chemical structure recognition tools on two data sets from the European Patent Office's Espacenet database: one with general organic chemistry structures and another with per- and polyfluoroalkyl substances (PFAS). While the tools performed well on standard organic structures, they struggled with PFAS structures, failing to correctly interpret 26 out of 43 unique PFAS structures. This highlights the need for quality control, confidence scoring, and expert review to ensure accurate automated screening of patents and strengthen early warning systems for hazardous chemicals.",
  "summary": "Patents are far more than legal documents protecting intellectual property. They are a vast source of scientific information about chemicals under development, often years before they appear in products or in the environment. That makes patents a valuable data source for early warning systems (EWS), which authorities use to identify potentially hazardous chemicals before they become a threat to…",
  "key_points": [
    "AI system converts patent images to machine-readable chemical structures",
    "Researchers at Umeå University developed automated workflow",
    "PFAS structures posed challenge, 26 out of 43 misinterpreted"
  ],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}