{
  "id": 6358615,
  "title": "IIT Indore Scientists Crack Open AI’s ‘Black Box’, Find Machines Can Learn Laws Of Chaos",
  "url": "https://urgent.news/2026/09/09/iit-indore-scientists-crack-open-ais-black-box-find-machines-can",
  "topic": "ai",
  "section": "AI",
  "published": "2026-09-09T00:30:00.000Z",
  "source": {
    "name": "Free Press Journal",
    "slug": "free-press-journal",
    "url": "https://www.freepressjournal.in/indore/iit-indore-scientists-crack-open-ais-black-box-find-machines-can-learn-laws-of-chaos"
  },
  "original_language": "en",
  "account": "Indore, Madhya Pradesh: Researchers at IIT Indore have uncovered evidence suggesting that artificial intelligence (AI) can glean physical laws governing sudden shifts in complex systems rather than merely recalling past data. This discovery could make AI more transparent and trustworthy. Led by Prof Sarika Jalan, the study utilized a machine-learning method called Reservoir Computing to analyze how an AI model anticipates critical transitions, or tipping points. The research forms part of PhD candidate Dishant Sisodia's doctoral work. A key hurdle in AI technology is its \"black box\" nature: while AI can make accurate predictions, understanding how it arrives at those conclusions is often challenging. To bridge this gap, the IIT Indore team devised physics-based tools to compare the AI's internal workings with the physical systems it was designed to emulate. The researchers discovered that the AI mirrored real-world systems, including minor statistical features appearing fractions of a second before a crisis. This indicates that AI was not merely recalling patterns but learning fundamental dynamical rules. Similar behavior was observed across multiple chaotic systems, bolstering the evidence. IIT Indore Director Prof Suhas Joshi emphasized that the research could enhance AI's reliability and explainability, as AI becomes more prevalent in daily life. Jalan highlighted that global research combining dynamical systems and machine learning is still nascent, particularly in India. By merging their expertise in chaos theory and non-linear dynamics with cutting-edge artificial intelligence, the team aims to create efficient, predictable, and controlled AI systems. The research's potential applications include improving early-warning systems for climate tipping points, financial market crashes, and medical events like epileptic seizures, as well as enabling scientists to use physics to understand machine-learning models and transparent AI for better prediction of complex systems.",
  "summary": "Indore (Madhya Pradesh): IIT Indore scientists have found evidence that artificial intelligence (AI) can learn physical laws governing sudden changes in complex systems rather than merely memorising past patterns, potentially making AI more transparent and trustworthy. The study, led by Prof Sarika Jalan, used a machine-learning technique called Reservoir Computing to examine how an AI model…",
  "key_points": [],
  "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."
}