Dutch Watchdog Calls for Swift AI Standardization
The urgency for standardization in artificial intelligence systems under the EU’s AI Act has been emphasized by the Dutch privacy watchdog, Autoriteit Persoonsgegevens (AP). The director of coordination and supervision on algorithms at the agency, Sven Stevenson, highlighted the necessity for accelerating the standardization process to provide certainty for companies aiming to demonstrate compliance. The European Commission tasked CEN-CELENEC and ETSI with developing these standards in May of the previous year, yet progress remains slow. With the AI Act’s provisions gradually coming into effect, including rules for General Purpose AI models by August next year, there is a pressing need for faster action.
The AP, functioning as the data protection authority in the Netherlands, is poised to share the responsibility of ensuring compliance with the AI Act alongside other agencies, such as the RDI. With a dedicated team of around 20 individuals focusing on AI, the AP has already taken action against breaches of the General Data Protection Regulation (GDPR), such as fining Clearview AI €30.5 million for illegal data practices. The AI Act is set to complement the GDPR by focusing on product safety, ensuring consistent regulation across EU member states.
To aid businesses in adapting to the forthcoming AI Act, the European Commission has initiated an AI Pact, facilitating workshops and joint commitments. Concurrently, the AP, in collaboration with the RDI and the Economic Affairs Ministry, is developing a sandbox and pilot project scheduled for 2026. This initiative aims to provide clarity and guidance for companies on compliance with the AI Act, particularly for AI systems with broader impacts.
Moreover, the Dutch government has introduced a public algorithm register since December 2022, aiming to ensure transparency and accountability in governmental algorithm usage. This measure seeks to legally evaluate algorithms for discrimination and arbitrariness, thereby enhancing transparency and explicability in algorithmic outcomes.
Source: Time’s running out on AI standardisation process, Dutch watchdog warns
