Brussels, Belgium (Enmaeya News) — Starting Aug. 2, new provisions of the European Union’s AI Act will come into force, marking a major step in regulating artificial intelligence across member states.
The changes include the appointment of national authorities to oversee AI compliance, the introduction of penalties for violations, and rules specifically targeting general purpose AI systems like ChatGPT and Google’s Gemini.
By the deadline, EU member states must notify the European Commission which market surveillance authorities will monitor businesses under the AI Act. However, many countries are behind schedule, with experts warning delays could hamper effective oversight.
“National authorities must be competent and properly resourced to oversee the broad range of risks posed by AI systems, including threats to fundamental rights,” said Laura Lazaro Cabrera, program director at the Centre for Democracy and Technology.
Setting up regulators is a complex task, involving coordination across AI, cybersecurity, data protection, and sector-specific expertise. Artur Bogucki, a researcher at the Centre for European Policy Studies, noted the challenge of aligning these efforts with existing rules like GDPR and the Digital Services Act.
Penalties for AI Act breaches will also be introduced. Companies may face fines up to €35 million or 7% of their global turnover, whichever is higher. Member states will determine how to implement these fines, with smaller firms facing lower penalties.
“There will be variability on how member states fine for non-compliance,” Lazaro Cabrera said. “But authorities have jurisdiction over any product entering the EU market, and fines are just one enforcement tool.”
Rules on general purpose AI systems (GPAI) come into effect as well. In July, the European Commission published a voluntary Code of Practice on GPAI, covering transparency, copyright, and security. Providers who sign the code are seen as compliant, but all must still follow the AI Act’s requirements.
US tech giant Meta recently declined to sign, citing concerns the rules could stifle innovation, while companies like Google and OpenAI have signed on.
Products already on the market before Aug. 2 have a two-year grace period to comply, but any new AI tools launched after must meet the rules immediately. The AI Act will be fully applicable starting Aug. 2, 2027.
The EU’s phased rollout of AI regulation reflects the growing complexity of balancing innovation with protecting fundamental rights and public safety.
