Эксперт заявил о необходимости чистых данных и ограничении использования синтетических в системах ИИ

Clean and diverse data are essential for reliable artificial intelligence (AI) systems, while synthetic data should be used within clearly defined limits, said AI technology expert Nilanjan Dey, Professor, Techno International New Town. Speaking at the international conference on Quantum Enhanced Sustainable Technologies for AI Systems (QUEST-AI), organised by Andhra University Engineering College, Mr. Dey said machine-learning algorithms depended on data and that poor-quality data could produce flawed results. He said the focus should be on clean and diverse data rather than merely on “big data”, with greater attention also needed on “small data” where data availability was limited. Mr. Dey said synthetic data could help when real-world data were scarce, but excessive dependence on it could increase errors in large language models (LLMs). He said its use should have clear limits. On the concluding day, Thursday, BARC senior scientist S. Praveen Kumar said quantum technology and AI were likely to work together in future. He said data centres would face major electricity requirements and suggested nuclear power as an option. He also said future facilities could become “quantum-enhanced AI data centres”. Greg Skulmoski of Bond University, Australia, stressed taking university research into real-life applications and adding business value to research. NVSN Sharma of NIT-Warangal spoke on Terahertz technology and its applications in defence and healthcare. Published - September 03, 2026 08:57 pm IST