{"input_url":"https://news.rpi.edu/2026/07/24/ai-growing-pains-reach-scientific-labs-new-rpi-study-finds","gnews_decode":"https://news.rpi.edu/2026/07/24/ai-growing-pains-reach-scientific-labs-new-rpi-study-finds","http_status":200,"final_url":"https://news.rpi.edu/2026/07/24/ai-growing-pains-reach-scientific-labs-new-rpi-study-finds","html_length":16812,"trafilatura_result":"ok","text_length":3484,"text_preview":"July 24, 2026\nResearchers at Rensselaer Polytechnic Institute (RPI) have found that today's leading artificial intelligence tools for predicting protein structures routinely generate results that are physically and chemically impossible, exposing critical blind spots in how AI is being applied acros","verdict":"OK"}