GenerativeAIPublicEN

Elevated Public Service Quality GenAI has the potential to elevate citizen experiences and boost overall satisfaction with public services through the customization of service delivery and expanded availability. It can be used effectively to optimize citizens’ experiences across various public sectors, including healthcare, education, social welfare, and legal assistance. Generative Artificial Intelligence Guidelines For Public When dealing with GenAI tools, developers and users should consistently adhere to the AI ethics and GenAI guidelines during all phases of the system’s lifecycle to harness their benefits while mitigating risks. The AI ethics principles developed by SDAIA apply to all segments of society in the Kingdom of Saudi Arabia. These include, but are not limited to, public, private, and non-profit entities, researchers, employees in both the public and private sectors, and consumers. The principles are designed to be applicable to all AI tools, not just GenAI, and are listed below: Fairness Requires stakeholders to take necessary actions to eliminate bias, discrimination, or stigma- tization of individuals, communities, or groups in the design, data, development, deployment, and use of GenAI systems. Given that GenAI tools have the capability to generate content that may be discriminatory or lack representativeness. To ensure consistent systems that are based on fairness and inclusiveness, GenAI systems should be trained on data that are cleansed from bias and is representative of affected minority groups. As such, developers and users should: • Carefully test GenAI models to insure no bias has been embedded in the codes or algorithms and Ensure that the data used to train the tool are cleansed from bias and is representative of affected minority groups. • Make efforts to gain a thorough understanding of the data used to train the tool – this un- derstanding should encompass the data’s origin, its contents, and how it was selected and prepared. • Enhance knowledge on bias, diversity, inclusion, anti-racism, values, and ethics – this knowl- edge will improve their ability to recognize biased or discriminatory content. 4. Principles of Generative AI Generative Artificial Intelligence Guidelines For Public Reliability & Safety Ensures that GenAI systems adhere to set specifications and that they behave as designers intend. Reliability is a measure of consistency and provides confidence in how robust a system is. On the other hand, safety is a measure of how the AI system does not pose a risk of harm or danger to society and individuals. A reliable and safe working system should have built-in mechanisms to prevent harm. As such, developers and users should: • Design and develop GenAI system that can withstand the uncertainty, instability, and volatil- ity that it might encounter. • Establish a set of standards and protocols for assessing the reliability of a GenAI system to secure the safety of the system’s algorithm and data output. • Predefined triggers/alerts should be in place based on system risks, GenAI systems should trigger human oversight. • System trigger should be assigned to the appropriate stakeholder. These triggers/alerts can be defined as part of the risk mitigation or disaster recovery procedure and may need human oversight. • Identify AI-generated content to ensure users are alert to potential reliability issues and can check it against other sources. • Encourage users to complement AI-generated content with information from reputable sources to ensure accuracy. • Strive to comprehend the quality and origins of the training data employed by the AI system to enhance content reliability. • Scrutinize content for factual accuracy and contextual relevance to prevent the dissemina- tion of erroneous information. Generative Artificial Intelligence Guidelines For Public