The China Academy of Information and Communications Technology has launched the first batch of testing and evaluation work for trusted internet intelligent agents

The China Academy of Information and Communications Technology announced that it will focus on core issues of user concern such as "functionality trustworthiness," "permission reliability," "operational transparency," and "behavioral intervention." It has officially launched the testing and evaluation work for trustworthy internet intelligent agents. The testing system revolves around three major categories: "capability authenticity," "permission reliability," and "behavioral controllability," with a total of 16 indicators and over 70 sub-items. Evaluations will be conducted based on the classification of internet intelligent agent service forms. The first batch of testing and evaluation officially starts today, with assessments carried out according to two forms and three types of internet intelligent agents. Material review, technical testing, expert evaluation, and other work are expected to take place from March to April 2026, with the final results to be released at the relevant industry conference in May 2026