According to Foresight News, Illia Polosukhin, co-founder of NEAR Protocol, spoke at the '2024 Hong Kong Web3 Carnival', advocating for the integration of Web3 and AI as a solution to the current closed-source language models. He pointed out that most AI models are closed-source, leading to a lack of transparency about the data used for training, which poses potential risks.
There is a need for open-source models, but several obstacles exist, such as regulatory restrictions on data access and insufficient financial incentives following open-sourcing. Polosukhin suggested that AI language models need to establish direct contact with users. The optimal solution, he proposed, is to use cryptography and on-chain reputation to address these issues, enabling the creation of competitive models within the Web3 framework.