EU's New AI Act: Landmark Rules for the Future of Digitalisation

The European Union is at the peak of establishing its first-ever AI legislation model in June 2024, becoming the world's first union to enforce such a cutting-edge Act. It has proposed a new set of rules to regulate technological issues, high risks, transparency requirements and the development of safe and trustworthy AI systems. This innovative AI regulation follows a 'risk-based' approach; the higher the risk of harming society, the stricter the rules. 

The AI Act, which was officially announced earlier this year, is designed to protect EU citizens, provides exceptions for diverse systems such as defence, military, and research purposes. It strikes a balance by prohibiting real-time biometric surveillance in public arenas, except for rare cases like terrorism, or the search for suspects involved in severe offences, ensuring privacy and security.

Categories of AI Systems Based on Risk

Based on risk, AI is categorised into new, lawful types:

• AI systems showing limited risk could be subjected to lighter transparency and accountability. 

• In contrast, high-risk AI systems could be approved but are subject to different requirements and obligations to gain entry into the EU market. 

• Due to unacceptable risk, the EU will block Social Achievement and Cognitive Behavioural Manipulation AI systems. 

• The new AI Act of the EU also banned predictive profiling-based policy systems that use biometric data to classify people according to race, sexual orientation and religion. 

Penalties and Measures for the EU AI Act

A fine of up to €35m or 7% of an organisation's global annual turnover awaits those who violate the AI Act. At the same time, start-ups & medium-sized companies are subjected to a pay fine according to the Administrative Fine Program.  The AI Act aims to provide innovative, evidence-based legal and regulatory learning and frameworks. AI Regulatory Sandboxes, by new law, enable a controlled environment for developing and verifying new AI systems in the real world. 

The Future of AI Legislation

So, what's the next step of the EU's AI act? Is it more engaged towards digitising the government or creating innovative safety laws and regulations?

Consider the 'Sao Paulo Guidelines' as a roadmap to a secure digital future, crafted with the principles of transparency, equality, and human rights at its core. These guidelines represent an ongoing journey towards the effective digitalization of government, which requires the active involvement of all technical and legislative bodies, collective efforts, and shared values for an authentically safe digital future.