The new service will validate "declared AI" agents such as ChatGPT, Claude and Perplexity deemed a proxy for human consumers, but will filter "undeclared, potentially malicious AI bots." ...