Google Research Identifies Recall Limitations in Frontier AI Models

A study by Google Research highlights significant recall constraints in advanced AI models such as GPT-5 and Gemini-3. The findings suggest that enhancing internal recall mechanisms could improve factual accuracy and reduce the dependency on massive datasets and external retrieval systems.