CIDOB Opinion 784 highlights the growing importance of AI governance and the European Union's pioneering role in creating a world-first AI regulation. The EU has developed a comprehensive framework that categorizes AI applications based on risk levels, aiming to limit potential dangers. This includes prohibitions on emotional recognition systems, intentional manipulation for influencing behavior, and those exploiting specific groups' vulnerabilities.
The regulation imposes obligations and restrictions on AI usage according to their risk, with penalties and fines for non-compliance. The EU has become a leader in defining legal frameworks for AI, yet criticism arises over its focus on mitigating risks rather than addressing inherent technology issues. There is an emphasis on diagnosing problems rather than just finding solutions.
The document discusses the balance between the EU as regulator and user of automated decision-making systems in the public sector. It points out the necessity for transparency and the recognition of AI's discriminatory impacts, which have been perpetuated by biases in algorithms affecting predictions, recommendations, and automated decisions. The author argues for a shift from regulatory solutions to address issues to understanding the origins and causes of AI's current challenges and the economy it supports.
In this context, the debate around AI governance is intertwined with the geopolitical race for technological innovation leadership. The EU is seen as a key player in this race, relying on its regulatory credibility. The debate is expected to be crucial in 2024, despite disagreements on AI definition among major actors. The focus shifts not only on whether to regulate AI but also on who should do so. The summit on AI security held in November by British Prime Minister Rishi Sunak exemplifies the power of dominant tech companies in shaping regulatory boundaries, potentially leading to a new form of colonialism.