As AI adoption accelerates, its applications in sensitive infrastructure—such as energy systems, industrial IoT, and critical facilities—are both promising and challenging,The application of AI in these critical domains can enhance operational efficiency, safety, and sustainability, but it also raises concerns about the reliability and trustworthiness of the systems that support vital societal functions. A very important area of research that has been rapidly evolving and being linked to trustworthiness of AI, as experts in multiple fields especially sensitive infrastructure have been hesitant into trusting AI, which is AI trustworthiness has become a trendy AI research area. This session aims to explore cutting edge research, challenges, and best practices to ensure AI reliability and trustworthiness in these critical domains.

Discussions will emphasize mitigating bias in all its different aspects(bias aspect, algorithm aspect), ensuring algorithmic fairness, and implementing ethical practices. Additionally, the role of AI standards and international initiatives (e.g., the EU AI Act) in shaping the future of AI governance will be examined, with special attention to how nations, including Tunisia, can contribute to these efforts.

The session will also focus on practical aspects, including optimizing AI models for resource-constrained environments, developing explainable AI frameworks(trustworthy dataset for AI such as PANDORA project ), furthermore, interdisciplinary collaboration between software engineers, data scientists, and domain experts will be essential in addressing issues such as algorithmic unfairness, lack of diversity in AI development teams, and the potential for exclusionary practices that fosters interdisciplinary collaboration (e.g., with software engineers and data scientists) to address challenges like bias, algorithmic unfairness, and lack of diversity among AI developers. By highlighting both theoretical insights and real-world case studies, the session seeks to advance understanding of how reliable AI systems can be designed and deployed in sensitive sectors while ensuring ethical and fair outcomes. Through a deeper exploration of these challenges and opportunities, participants will gain a comprehensive perspective on building AI that fosters trust and fairness in mission-critical applications.

Key Themes Covered in the Session

1. The future of AI in critical sectors: Challenges and opportunities
2. Trustworthiness and reliability of AI: Tackling bias, algorithmic unfairness, and diversity
3. Role of international standards (e.g., EU AI Act, NIST) and Tunisia’s potential contributions
4. Practical perspectives on deploying AI in resource-constrained environments
5. Collaborative roles of software engineers and data scientists in trustworthy AI.