Chainlink implements a new strategy to solve a major problem in artificial intelligence: hallucinating AI systems. When large language models incorrectly interpret data or generate incorrect new data, the consequences can be expensive, especially in the financing. Instead of trusting a single AI model, Chainlink now uses a multi-model approach, using AI systems from OpenAi, Google and Anthropic.
Laurence MoroneyA Chainlink consultant and former head of AI at Google explained that the use of multiple AI models instead of just one reduces the error percentage. Each AI model is asked separately to analyze the same financial data. The system stores verified data on the blockchain, making it transparent, unchangeable and safe. This method based on consensus prevents financial data from being damaged by incorrect information and increases the reliability of data generated by AI.
Chainlink’s approach is intended to change this by reducing manual data verification and increasing financial accuracy. In a recent collaboration with leading financial institutions, including UBS, Franklin Templeton, Wellington Management, Vontobel and Sygnum Bank, Chainlink tested this AI-driven blockchain system. The results were promising, which demonstrated a reduction in errors and inefficiencies in financial data.
Image: Freepik
Designed by Freepik
