Diritto intelligente – Issue N. 6

The EU Commission has released much-anticipated guidelines clarifying which technologies qualify as “AI systems” under the AI Act, offering businesses a clearer framework for distinguishing between regulated AI solutions and those beyond the Act’s scope.

This announcement arrives just as the long-awaited provisions on prohibited practices of the EU AI Act take effect. The bar is set high: certain applications deemed “unacceptably risky” are now off-limits. Organizations are now required to re-evaluate any initiatives that may infringe upon core principles, including social scoring or other invasive data practices.

In parallel, privacy professionals are focusing on more efficient and transparent ways to uphold data subjects’ rights. Retraining AI models to accommodate erasure requests remains a significant challenge, but stakeholders are steadily advancing techniques for “unlearning” personal data without compromising system performance.

Legal design methodologies also feature prominently this month. Through more intuitive drafting and visualization, even the most complex contractual or regulatory texts can be made accessible, ultimately supporting both compliance and user engagement.

Finally, in-house teams planning to deploy AI will find practical guidance on aligning technological investments with overarching business objectives. By ensuring robust governance, adequate risk mitigation, and rigorous data controls, organizations can embrace AI’s transformative potential while respecting emerging legal and ethical standards.

Enjoy this month’s insights and remember: an informed approach will keep you ahead of the AI curve – no algorithmic troubles required.