Last month, the Ministry of Transport, Communications, and Information Technology (MTCIT) issued Oman’s first policy for governing the use of artificial intelligence (AI) systems. This policy was issued in the form of a circular, and therefore it is not a strictly binding legal document published in the Official Gazette.
We previously wrote a blog about this policy in August of last year, when it was still at the public consultation stage. Now that the AI governance framework has officially been issued, it’s worth noting that, as expected, the MTCIT has chosen to release it as a non-legally binding policy. While this means that it does not impose fines or include enforcement mechanisms, this approach has its advantages as it allows for flexibility and encourages voluntary compliance without imposing serious restrictions on the adoption and experimentation of this cutting edge technology.
Furthermore, as predicted in our blog on the implications of the EU AI Act for Oman, the EU’s regulatory approach had an influence on Oman’s approach in governing AI. While the MTCIT’s AI governance framework is non-legally binding (unlike the legally binding EU AI Act), it borrows several concepts from the EU AI Act, particularly in regard to the distinction between the obligations of AI system providers and AI system deployers.
The policy sets a general framework of technical controls and ethical principles that both public and private sector entities need to take into account to when developing or deploying AI systems. Those ethical principles include considering humanity and societal well-being, inclusivity, fairness, and responsibility when developing or deploying AI systems. This policy also includes provisions on periodic evaluation of AI systems for accuracy, bias, and harmful content; implementing human oversight for crucial AI decisions and their societal impact; ensuring deployed AI systems provide tangible benefits; adhering to data protection laws, restricting data use with robust security and cybersecurity best practices, and documenting AI processes for transparency.
The policy has already entered into force on 9 April 2025. You can read it in full in English on the link below: