Understanding the EU Guidelines on Prohibited AI Practices under the AI Act

Executive Summary

The European Commission has published guidelines clarifying prohibited artificial intelligence (AI) practices as defined in the EU's Artificial Intelligence Act (AI Act).

These guidelines aim to ensure consistent application and understanding of the Act across the European Union (EU), safeguarding innovation while prioritising health, safety, and fundamental rights.

Though non-binding, they provide critical insights into prohibited AI activities, such as harmful manipulation and social scoring, offering practical examples for stakeholders.

This blog explores the implications of these guidelines for insurance and financial sectors, alongside recommendations for effective risk management.

Prohibited AI Practices and Their Rationale?

The AI Act identifies specific AI practices that are outright prohibited due to their potential to harm individuals or society. These include:?

• Harmful manipulation of individuals' behaviour that exploits vulnerabilities.?
• Social scoring systems, which evaluate individuals based on their social behaviour or characteristics.?
• Real-time remote biometric identification in public spaces, except in narrowly defined circumstances such as law enforcement emergencies.?

These activities are banned to protect fundamental rights, prevent discrimination, and mitigate risks to personal autonomy and democratic values.

Scope of Application and Sectoral Implications

The guidelines apply across all EU member states, ensuring uniformity in enforcement. For life insurance and pension products, this means heightened scrutiny over any AI systems used for customer profiling or risk assessment to ensure they do not involve prohibited practices like social scoring. For non-life insurance and health insurance sectors, companies must ensure compliance when deploying AI tools for claims processing or fraud detection, avoiding manipulative or discriminatory algorithms.

Implications of Non-Compliance and Timeline

Breaching these guidelines could result in significant penalties under the AI Act, including fines of up to €30 million or 6% of global annual turnover. While the guidelines are currently in draft form, companies should prepare for their formal adoption expected later this year. Early alignment with these principles will minimise compliance risks.

Recommendations for Risk Management

To navigate these requirements effectively:?

1. Conduct comprehensive audits of existing AI systems to identify potential risks.?
2. Develop robust governance frameworks to ensure ethical AI deployment.?
3. Train teams on the ethical use of AI and compliance with EU regulations.?
4. Engage legal counsel to interpret the guidelines in context with your operations.?

?

For further information or tailored advice on managing these changes, please contact us at [email protected] or book an appointment to discuss.

?

Citations:

[1] https://digital-strategy.ec.europa.eu/en/library/commission-publishes-guidelines-prohibited-artificial-intelligence-ai-practices-defined-ai-act.