Lead
The European Union has entered a new phase of its Artificial Intelligence Act, with transparency requirements now in effect and regulators gaining direct enforcement powers. The legislation, adopted in 2024, is widely described as the world's first comprehensive law on AI, and its rollout is being watched globally as a potential model for AI governance.
Coverage comparison
Reports on the development vary in emphasis. Al Jazeera highlighted the delay in high-risk obligations and the controversy surrounding it, quoting critics who argue the postponement could have serious consequences for vulnerable groups. RFI — English focused on the comprehensive nature of the framework and its consumer protections, framing the Act as a landmark effort to regulate rapidly evolving technology. South China Morning Post emphasized the enforcement aspects, describing the regulations as among the most aggressive the tech sector has faced, and quoted the EU's tech sovereignty chief on the significance of the step.
Key claims
New transparency rules take effect
From August 2, Article 50 of the AI Act applies, requiring AI systems that interact directly with people to disclose that users are dealing with AI, unless it is obvious from the context. Providers of AI systems that generate or manipulate content must ensure synthetic content is identifiable through machine-readable markings. Systems that recognize emotions or categorize people using biometric data must inform individuals that such processing is taking place, with certain law enforcement activities exempt. Violations can result in fines up to 15 million euros ($17.3 million) or 3 percent of global turnover, according to Al Jazeera.
The rules also require AI-generated content to carry electronic markings, and those publishing deepfakes must disclose that material has been altered or artificially created, as reported by RFI — English.
Enforcement powers and new team
South China Morning Post reported that the European Union rolled out a new team to regulate AI companies worldwide, and that from the same date, the European Commission gained the power to enforce obligations directly, including conducting investigations and issuing fines. The Commission stated that new regulations include "systemic risks" posed by AI, such as "chemical, biological, radiological and nuclear incidents, loss of control, cyber offence, harmful manipulation and threats to fundamental rights".
Henna Virkkunen, the EU chief for tech sovereignty, said on Friday: "As enforcement begins, we are taking an important step towards AI that people and businesses can understand and trust, and whose benefits are shared widely across our society."
Delayed high-risk obligations
While the transparency provisions are now in effect, the AI Act's high-risk obligations are delayed and will apply from December 2027, as reported by Al Jazeera. This delay has drawn criticism from some quarters. Critics argue that postponing these obligations could increase surveillance and discrimination, and even result in asylum claims being unlawfully rejected based on personal characteristics or racialised suspicion, according to Al Jazeera. This claim is carried by a single outlet and has not been repeated in other reports available.
Global influence
Both Al Jazeera and RFI — English noted the potential for the AI Act to influence global AI governance, drawing parallels with the GDPR, which became a global benchmark for privacy practices. Al Jazeera specifically mentioned the so-called "Brussels effect," where EU regulations often become de facto global standards. Whether the AI Act will prove as influential for AI governance remains an open question.
Perspectives
- EU officials: Henna Virkkunen, the EU chief for tech sovereignty, framed the enforcement as a positive step toward trustworthy AI, emphasizing benefits for society.
- Critics: According to Al Jazeera, critics argue that the delay in high-risk obligations could lead to increased surveillance and discrimination, with potential harms to vulnerable groups.