Introduction to the Annual Conference of the Tunisian Association of Investors in Capital (ATIC)
The Ministry of Information and Communication Technologies (TIC), represented by Tarak Triki, Director General of Smart Capital, commended the consistent commitment of the ATIC during the opening of its 2026 annual conference. The ministry highlighted the association's growing importance as a key player in Tunisia's financial ecosystem.
The Relevance of Artificial Intelligence and Sustainable Investment
The ministry also emphasized the significance of the theme chosen for the 2026 edition: artificial intelligence (AI) and sustainable investment. According to the ministry, this theme aligns perfectly with the current context of accelerated transformation, where AI is no longer seen as a theoretical promise, but as a concrete professional reality that deeply transforms the competitiveness of companies and decision-making processes.
The Need for AI Adoption and Governance
The ministry stated that the debate is no longer about the need to adopt AI, but rather about the speed of its integration, governance modalities, and deployment strategies. The ministry emphasized Tunisia's willingness to be an engaged actor in this technological revolution, rather than a mere spectator, and to accelerate its efforts to align with global mutations.
The Importance of Coordination and Action
The ministry stressed the need to strengthen coordination among all stakeholders and intensify existing actions. To conclude, the ministry announced that it would meet with ATIC board members after the conference to transmit recommendations and conclusions from the day's discussions.
Artificial Intelligence and Trust
During the conference, researcher and AI expert Nozha Boujemaa discussed the emergence of "trustworthy AI." She called for a paradigm shift in evaluating and governing AI systems, emphasizing the need for objective and measurable responsibility. Boujemaa argued that traditional approaches to AI ethics, often based on general principles and context, have reached their limits.
Performance vs. Trust
Boujemaa distinguished between performance and trust, stating that performance refers to a model's efficiency in controlled conditions, while trust refers to the system's robustness and ability to maintain quality in varied, changing, and sometimes unpredictable environments. She emphasized that robustness should be considered a quantifiable indicator of reliability.
Risks Associated with Insufficiently Mastered AI
Boujemaa highlighted the business risks associated with insufficiently mastered AI, including reputation risks that can impact customer trust and economic viability. She also warned about the potential effects on mental health when digital systems provide inadequate or harmful content. Boujemaa stressed that AI plays a growing role in shaping public opinion, raising questions about democratic sovereignty.
Evaluating and Managing AI Risks
Boujemaa discussed the need for frameworks to evaluate and manage AI risks, citing the National Institute of Standards and Technology's (NIST) reference framework, which identifies three levels of potential impact: harm to individuals, damage to businesses, and disruption to economic, social, or environmental ecosystems.
Testing and Anticipating Vulnerabilities
Boujemaa presented the "red teaming" approach, which involves actively testing AI models for vulnerabilities, biases, and undesirable behaviors before deployment. She also introduced the "Value by Design" approach, which integrates ethical, social, and economic values into algorithm design as structuring constraints.
Mastering the AI Life Cycle
In conclusion, Boujemaa emphasized the need to master the AI life cycle, from design to deployment, to ensure trustworthy and reliable AI systems.
