Google has made fixes to its artificial intelligence (AI) system that generates summaries for search results after it produced inaccurate information about certain topics. The company acknowledged the issue and said it has made adjustments to its systems to address the inaccuracies. The AI system, known as a large language model, was designed to generate readable summaries from web pages. However, it sometimes produced incorrect information, particularly on topics like recent events or biographies. Google said it has retrained the model using a different approach to reduce such hallucinations. The company emphasized that the AI-generated summaries are a supplement to search results and not a substitute for authoritative sources. Google plans to continue improving the system and is exploring ways to make it more transparent when summaries are generated by AI.