The European Union (EU) has recently reaffirmed its pledge to follow through with its timeline. It’s continuing to make progress on rollout of its landmark AI legislation, the AI Act. This decision comes on the heels of extreme pressure. You wouldn’t know it judging by the over 100 technology companies that are lobbying to weaken or delay the rollout of these rules. The announcement, reported by Reuters, reinforces the EU‘s commitment to regulate artificial intelligence through a robust framework.
The AI Act stands as one of the most impactful achievements in the global regulation of AI applications, particularly in its risk-based approach. This legislation defines a risk-based framework, classifying specific AI applications based on their level of risk. The most dangerous applications, deemed as posing “unacceptable risk,” are prohibited entirely. Cognitive behavioral manipulation and social scoring are clear examples of high-risk applications. Both of these practices are steeped in troubling ethical issues.
After nearly two years of debate, the EU has passed its AI regulations. Hundreds of thousands of letters and public comments have called attention to the harmful implications of the AI Act. Stakeholders from all industries and sectors are contributing to the discussion. Yet they call for a measured, nimble approach that encourages innovation but maintains strong safety and ethical guidelines.
The timeline for implementation of the AI Act is a crucial component of the EU’s overall plan. It intends to establish an overall regulatory landscape for AI technologies. By sticking to this schedule, the EU aims to lead in global AI governance, setting standards that may influence international practices.
The EU’s move to enshrine risk-based regulations in law further emphasizes the bloc’s commitment to tackling potential dangers posed by artificial intelligence. That is the explicit goal of this regulatory framework — to lower risks but more effective. It further encourages developing and using AI technologies responsibly.
