Larry Ellison highlights core data flaw affecting all AI models

Larry Ellison said the biggest challenge facing AI models is unreliable and outdated data, which leads to confident but incorrect responses. The Oracle cofounder argued that systems like ChatGPT, Gemini, Grok, and Llama depend heavily on training quality rather than scale alone. He emphasized the need for continuously updated, verified enterprise data to improve accuracy, trust, and real-world usefulness of artificial intelligence systems globally today.