Artificial Intelligence (AI) is defined as a branch of computer science concerned with creating technological systems that exhibit behaviour considered intelligent, simulating human intelligence. It aims to develop systems capable of thinking. AI encompasses various technologies such as machine learning, artificial neural networks, natural language processing, robotics, and others. These technologies work to represent knowledge, learn from data, and make decisions.

AI has strongly entered all areas of life, and the future will bring diverse applications of AI. However, the use of AI technologies and applications must be guided by legal and ethical principles that must be respected and adhered thereto, so that the benefits of AI are not lost and it does not become a threat to the security and stability of society and individuals.

Among the most important principles and rules that must be followed when using AI are fairness, accountability, and transparency principles which are emphasized in the Dubai Principles for Artificial Intelligence. Below is an explanation of each principle:

First: Fairness

Where it shall be provided by the requirements for achieving justice in AI systems that include in the following:

1. The data received by the system should represent the affected group as accurately as possible.
2. Algorithms must avoid operational bias.
3. Measures must be taken to assess any bias in datasets.
4. The fairness of general decisions must be demonstrable.

Second: Accountability

AI systems are accountable through the following:

1. Responsibility for AI outcomes does not lie with the system itself, but is divided  among those responsible for its design, development, and implementation.
2. Developers must make every effort to mitigate inherent risks in the systems.
3. AI systems must include procedures that allow users to challenge significant decisions.
4. AI systems should be overseen by diverse and expert teams.

Third: Transparency

Transparency refers to clarity and integrity in processes and decisions within the context of AI technology. It involves disclosing how algorithmic systems operate and the data used to make decisions. Transparency is achieved through:

1. Developers must build systems whose failures can be tracked and diagnosed.
2. Individuals must be informed of significant decisions made by AI.
3. Publishers of AI systems must be transparent regarding the algorithms and data they use, while maintaining intellectual property and privacy.

In our view, these principles must operate harmoniously, coherently, and in coordination to foster trust among individuals in AI systems and applications, encouraging positive engagement with them.

Dr. Ghassan Salim Arnous
Legal Consultant
Zayed Al Shamsi Law Firm and Legal Consultancy