AI Act: Europe Establishes Ethical Framework Amidst Rise of AI

In the face of the rapid rise of artificial intelligence, Europe is seeking to establish a robust regulatory framework with the AI Act. This regulation aims to ensure the ethical use of AI within the European Union. Initially scheduled for 2026, its rollout will now extend until 2028.

The AI Act imposes obligations on the training of models and protects intellectual property, which is crucial for the ethical use of AI. Mathieu Changeat, Chief Operating Officer of the Zaion Dydu group, emphasizes that the real challenge for Europe lies in the lack of funding compared to American and Chinese giants. Although European players like Mistral AI are emerging, the financial power remains insufficient to compete in the generative AI market.

The AI Act will introduce a compliance label that will enhance user trust. However, this label does not guarantee the absence of algorithmic hallucinations, which depend on the technical architecture of the models.

The quest for digital sovereignty is becoming a major issue for large companies and the public sector in Europe. These actors now favor local solutions, even if they are less efficient, in response to the current geopolitical situation.

However, a technological divide could widen. SMEs, with fewer resources, risk turning to standardized solutions, often American, due to a lack of means to keep pace with regulatory requirements.

The AI Act also addresses the governance of autonomous AI agents, requiring technical safeguards to prevent unforeseen behaviors. The growing distrust of users regarding the management of their personal data is pushing companies to anticipate transparency requirements. The cookie rejection rate is expected to reach nearly 40% by 2026.

The main "red flag" for a company remains the "black box" nature of certain AI models. The evolution of projects can render initially stable systems incompatible, which the AI Act could penalize.

Finally, despite fears of massive automation, AI is seen as a tool to enhance human capabilities rather than replace them. Users continue to prefer human interaction in complex situations. Since the AI Act is not a labor code, it will not have a direct impact on the protection of employees against automation.

By 2028, the AI Act is expected to act as a reassurance label rather than a constraint, fitting into a regulatory continuity with the GDPR. Mathieu's ultimate strategic advice is to focus on real business needs rather than succumbing to the trend effect of AI. It could catalyze a more ethical and sovereign AI, although financial and technological challenges persist.