Key Points
- The European Union introduces groundbreaking AI regulation, setting a global standard.
- The legislation places stringent controls on “unacceptable” AI uses, including social scoring and emotion interpretation.
- High-risk AI applications in education, hiring, and government services face new transparency obligations.
- AI companies like OpenAI are subject to mandatory disclosure requirements and labelling of AI-generated content.
- The EU’s rapid legislative process contrasts with the United States’ slower progress on AI governance.
The European Union (EU) has taken a pioneering step with its landmark law on artificial intelligence (AI), establishing itself as a frontrunner in AI regulation and leapfrogging the United States. This ambitious legislation, aimed at reshaping AI use in healthcare and policing, seeks to safeguard rights and maintain ethical standards.
EU Bans “Unacceptable” AI, Marks High Risk
Central to the EU’s regulatory framework are the outright bans on what it deems “unacceptable” uses of AI. This includes the prohibition of social scoring systems, the use of AI to deduce sensitive attributes such as race, political inclinations, or sexual orientation, and the deployment of AI for emotion interpretation in schools and workplaces. Additionally, the legislation stops certain automated profiling techniques to predict future criminal behaviour.
Equally significant is the designation of a “high-risk” category for AI applications, particularly those involved in education, employment, and government services. This category subjects AI to enhanced scrutiny, including transparency and accountability requirements.
OpenAI Faces New EU Disclosure Rules
The implications for AI companies under this new EU regulation are profound. Entities like OpenAI will now face stringent disclosure mandates, including the mandatory labelling of AI-generated content. Furthermore, deepfake technology’s potential for misinformation and manipulation necessitates its regulation, thus leading to its banning in certain contexts. This marks a pivotal shift towards greater transparency and responsibility in the AI sector.
EU’s AI Law: 2 Years to Enactment
Scheduled to take effect approximately two years from its approval date, the EU’s AI legislation highlights the bloc’s rapid and decisive response to the challenges and opportunities presented by AI advancements. Furthermore, stemming from a proposal introduced in 2021 and catalysed by the popularity of tools like OpenAI’s ChatGPT, this legislative initiative highlights the EU’s commitment to shaping a future where AI serves the public good within clearly defined ethical and legal boundaries.
Contrasting Dynamics: The EU and US on AI Regulation
While the EU charts a course towards comprehensive AI regulation, the United States lags in making comparable legislative strides. Despite efforts by figures such as Senate Majority Leader Chuck Schumer to prioritise AI legislation and Elon Musk suing OpenAI, the US has yet to achieve significant progress in this arena. This contrast underscores governments’ varying paces and priorities in addressing the complex landscape of AI technology and its societal impacts.
