Decentralization AI infrastructure Mira Network launches Proof of Agent, revealing the dangers of AI large model hallucinations.

According to Deep Tide TechFlow news, on April 27, the Decentralization AI infrastructure Mira Network announced the launch of an Agent based on OpenAI o3, which will provide false evidence and information sources based on user input, even if the user inputs a wrong statement.

This Agent aims to reveal the dangers posed by the current AI large model hallucinations, as AI can skillfully fabricate evidence, citations, and quotations.

As a Decentralization AI infrastructure, Mira Network is a middleware network specifically designed for the validation of AI LLMs. It is committed to building a reliable validation layer between users and underlying AI models, providing integrated services such as privacy protection, accuracy assurance, scalable design, and standardized API interfaces, to reduce the hallucinations produced by AI LLMs and promote the implementation of AI in various niche application scenarios.