The European Union (EU) has introduced a proposal for a regulation that lays down harmonised rules on artificial intelligence (AI), aiming to address potential risks and harm associated with AI technology. The regulation aligns with the EU’s commitment to preserving technological leadership and ensuring that citizens benefit from new technologies in line with Union values and fundamental rights. The proposal outlines various aspects, including the context, reasons, and objectives of the regulation, as well as harmonised rules, impact assessment, stakeholder consultation, and recommendations.

Context and objectives

The proposal recognises AI’s evolving nature and potential benefits across sectors, emphasising the need for a balanced approach to avoid negative consequences. It aims to regulate AI in specific areas such as finance, mobility, climate change, environment, and health, ensuring adherence to social values and fundamental human rights. The primary goal is to provide a consistent approach to deploying AI technologies within the EU.

Harmonised rules on AI

The regulation categorises high-risk AI systems, specifying requirements for notification, transparency, data governance, record-keeping, and conformity assessment. These rules are particularly applicable to AI systems in critical infrastructure, health, and workplace management. Additional provisions cover testing, privacy, and oversight measures.

Impact assessment

The regulation acknowledges the costs involved in providing human and financial resources for deployment and implementation. This resource allocation is essential for Member States’ supervisory arrangements to ensure compliance with the regulation.

Stakeholder consultation

Stakeholders participated in an online consultation, revealing a general consensus on the need for legislative action. While over 80% of business and industry representatives acknowledged legislative gaps, concerns were raised about potential duplication, conflicting obligations, and overregulation.

Evaluation and review

The Commission plans to review and evaluate the regulation five years after its entry into force, reporting findings to relevant EU bodies.

Recommendations

Several recommendations emerge from the proposal, including the need for supervisory mechanisms, regular reviews, adequate resource allocation, a technology-neutral regulatory framework, clear AI definitions, and designated supervisory authorities. The proposal also outlines the establishment of a system for registering high-risk AI applications in a public EU-wide database.

ESG issues

The proposal addresses critical Environmental, Social, and Governance (ESG) issues, emphasising the importance of respecting fundamental human rights, ensuring AI system safety in critical areas, and implementing monitoring and control measures to mitigate risks.

Conclusion

The proposed regulation aims to safeguard citizens’ fundamental rights, establish ethical frameworks for AI deployment, and minimise societal risks. It strives for a balance between AI benefits and accountability, transparency, and governance. Stakeholders can request updates to regulations to address evolving AI technologies and changing risks. Overall, the proposed regulation sets an ambitious framework for AI deployment while maintaining a commitment to fundamental human rights and ethical considerations.