المركز الوطني للتعليم الإلكتروني يصدر 23 معيار لاستخدام وكلاء الذكاء الاصطناعي في التعليم والتدريبSaudi Arabia's National eLearning Center issues 23 standards for using AI agents in education and training

Saudi Arabia's National eLearning Center has issued a document called "Standards for AI Agents in Education and Training", with 23 standards spread across 7 dimensions plus specific requirements for 8 areas of use, according to the Alyaum newspaper. The aim is to improve learning and service quality while controlling what agents are allowed to do and protecting the rights of the people they serve.

The document covers general and higher education, vocational and technical education and training, and lifelong learning, at government, private and non-profit bodies. It distinguishes three things: an automated workflow that follows preset steps, a tool that answers a request while a person decides the next step, and an agent that is given a goal, chooses some of the steps itself, carries them out, checks the results and adjusts its course within the permissions it has.

It sets three levels of agent autonomy: giving outputs and recommendations only, acting after a human check, or acting within defined permissions without approval for each step. It rates impact as limited, medium or high, and notes that an agent's recommendation can affect a learner's future even if the agent does not act on it itself.

The eight areas include teaching, assessment, guidance, learning content, research and administrative services. The document asks organisations for evidence of value and safety, not just fast or good-looking output. It is a guiding reference, and each organisation chooses the tools and procedures that suit its work.