Launched in 2015, the EU’s Digital Single Market Strategy aimed to foster the digital harmonization between the EU member states and contribute to economic growth, boosting jobs, competition, investment and innovation in the EU.
The EU AI Act characterizes a fundamental element of this strategy. By adopting the first general-purpose regulation of artificial intelligence in the world, Brussels sent a global message to all stakeholders, in the EU and abroad, that they need to pay attention to the AI discussion happening in Europe.
The EU AI Act achieves a delicate balancing act between the specifics, including generative AI, systemic models and computing power threshold, and its general risk-based approach. To do so, the act includes a tiered implementation over a three-year period and a flexible possibility to revise some of the more factual elements that would be prone to rapid obsolescence, such as updating the threshold of the floating point operations per second — a measurement of the performance of a computer for general-purpose AI models presumed to have high impact capabilities. At the same time, the plurality of stakeholders involved in the interpretation of the act and its interplay with other adopted, currently in discussion or yet-to-come regulations will require careful monitoring by the impacted players in the AI ecosystems.
(more…)
