Researchers develop new way to eliminate AI hallucinations
Researchers develop new way to eliminate AI hallucinations
Publish Date: 2026-05-28 08:06:00
Source Domain: www.binghamton.edu
Here is a summary of the article in an unordered list highlighting 5 key points:
-
Research on AI Chatbots in Medical Diagnosis: Chatbots powered by AI are increasingly used to help diagnose medical concerns, though they can generate false information and “hallucinations.”
-
Development of Verification Method: Researchers Ahmed Abdeen Hamed and Luis M. Rocha have developed a protocol to reduce false information or “hallucinations” from chatbots by using retrieval-augmented generation (RAG) to verify outputs against authoritative databases.
-
Seven LLMs Verification Experiment: Over 10,000 experiments with seven selected large language models showed 76.85% of answers were supported by at least four models, ensuring no unmatched or “hallucinated” terms.
-
Versatile Protocol Applications: The protocol is not only effective for medical applications but also has the potential to verify fabricated information in domains like legal research, academic citations, and historical facts.
-
Future Prospects: The innovation aims to increase confidence in large multiscale network models of disease and facilitate precision medicine, including the development of “digital twins” for healthcare, while also pushing the boundaries of knowledge verification in various fields.
