The EU AI Act

The EU AI Act is a comprehensive legal framework proposed by the European Commission that aims to regulate the development, deployment, and use of artificial intelligence (AI) technologies within the European Union. It is one of the first significant legislative efforts by any government or international body to create a coherent set of rules for AI, and it has the potential to set global standards for AI governance.

The EU AI Act entered into force on August 1, 2024. However, the enforcement of most of its provisions will begin on August 2, 2026.

Risk-Based Approach

The EU AI Act introduces a four-tiered, risk-based classification system for AI systems, categorizing them based on the level of risk they pose to individuals' rights and safety. These categories are:

1. Unacceptable Risk: AI systems that pose a clear threat to people's safety, livelihoods, or rights are prohibited. This includes:

• AI systems that use subliminal methods to distort a person’s behavior in a way that causes or is likely to cause physical or psychological harm.

• AI systems that exploit the vulnerabilities of specific groups (e.g., based on age, mental or physical disability) to distort their behavior in a way that could cause physical or psychological harm.

• Social scoring systems used by public authorities that lead to discriminatory outcomes.

• AI systems used for "real-time" biometric identification in public spaces by law enforcement (with some exceptions).

2. High-Risk AI: These systems are subject to strict regulations and oversight. They include AI used in critical areas such as:

• Critical infrastructure (e.g., transport systems) that could endanger lives.

• Education/training affecting access to opportunities (e.g., exam scoring).

• Safety-critical products (e.g., robot-assisted surgery).

• Employment (e.g., AI for recruitment processes).

• Essential services (e.g., credit scoring for loans).

• Law enforcement (e.g., evaluating evidence reliability).

• Migration/border control (e.g., automated visa processing).

• Justice and democratic processes (e.g., searching court rulings).

Obligations for high-risk AI before market deployment include:

• Risk assessment and mitigation.

• High-quality, non-discriminatory datasets.

• Logging for traceability.

• Comprehensive documentation for compliance.

• Clear information to users.

• Human oversight to minimize risks.

• Robustness, security, and accuracy.

3. Limited Risk AI: Limited risk relates to the lack of transparency in AI usage.The AI Act addresses this by requiring:

• Users to be informed when interacting with AI, like chatbots, allowing them to make informed choices.

• Providers to ensure AI-generated content is clearly identifiable.

• AI-generated text, audio, and video intended for public interest, including deepfakes, must be labeled as artificially created.

4. Minimal Risk AI: These includes systems like spam filters or AI video games. These are subject to voluntary codes of conduct rather than formal regulation.

Applicability to Hypatos

From a regulatory perspective, Hypatos’ AI solutions are best characterised as minimal risk AI under the EU AI Act. The system performs back-office automation by structuring and classifying existing data and does not generate synthetic content, materially alter original inputs, or make autonomous decisions producing legal or similarly significant effects for individuals. The output constitutes a technical transformation of pre-existing information into structured data.

Nevertheless, an argument could be made that certain functionalities fall within the limited risk category due to user interaction with AI-supported features. On a precautionary basis and in anticipation of evolving regulatory interpretation, we therefore align our practices with the transparency expectations typically associated with limited-risk systems, including clearly informing users when they interact with AI.

How Hypatos Approaches AI Compliance

At Hypatos, we recognize the importance of AI legislation that governs the responsible use of AI technologies. As developers of advanced AI-driven solutions for back-office automation, we are committed to ensuring that our AI systems align with the EU AI Act's requirements.

We are committed to staying informed about regulatory changes, ensuring compliance, and promoting responsible AI use for our customers. Our solutions help businesses automate repetitive tasks like invoice processing, document extraction, and data validation, freeing up valuable time for employees to focus on more strategic and creative tasks.

Conclusion

The EU AI Act is a pivotal step in shaping the future of AI regulation, and at Hypatos, we are committed to staying at the forefront of these developments. We believe that responsible AI is the key to unlocking the full potential of automation while ensuring trust, transparency, and compliance. By aligning our AI solutions with regulatory standards, we empower businesses to automate routine tasks, focus on strategic growth, and confidently navigate the evolving world of AI.

_{Edwards, L. (2022). Expert explainer: The EU AI Act - A summary of its significance and scope. Ada Lovelace Institute. https://www.adalovelaceinstitute.org/wp-content/uploads/2022/04/Expert-explainer-The-EU-AI-Act-11-April-2022.pdf}

‍