PANews reported on June 9th that, according to The Block, a report by IC3 researchers from top universities such as Cornell University and Carnegie Mellon University points out that claims about encryption technology solving AI's biggest challenges are often exaggerated, and that encryption has limited effectiveness in addressing trust and payment issues in AI. The study indicates that providing AI agents with encrypted wallets will not make them smarter or more resistant to human manipulation; the wallets only allow AI agents to automate transactions and access on-chain infrastructure without human approval.

Regarding the distinction between AI-generated and human-generated content, while blockchain is suitable for timestamping and registering specific digital content, it cannot determine how the content was created, requiring external tools for verification. If these external tools err and record the error on the chain, the error will be permanently retained. The role of blockchain is limited to providing highly complete registration records for certain types of content. Regarding algorithmic bias, research suggests that bias is unlikely to be resolved through decentralized AI because it originates in the training process and is typically mitigated by improving training or inference techniques. Decentralization cannot address the root cause of bias, and there is currently no evidence that transparency and increased participation reduce bias or improve model output.