# Understanding - EU AI ACT 2024 # EU AI ACT 2024 - Oerview ### **General Principles** - **Risk Categorization** - **Reference:** Article 6 – Classification of AI systems. - High-risk systems include those impacting education per Annex III. - **Transparency** - **Reference:** Article 52 – Transparency obligations for AI systems. - **Accountability** - **Reference:** Article 23 – Governance and accountability in high-risk systems. - **Human Oversight** - **Reference:** Article 14 – Human oversight requirements for AI systems. - **Ethical Framework** - **Reference:** Article 9 – Risk management system requirements. --- ### **Data Handling** - **Data Privacy** - **Reference:** Article 10 – Quality of datasets, aligned with GDPR (General Data Protection Regulation). - **Bias Mitigation** - **Reference:** Article 10 – Avoiding biases in datasets. - **Data Security** - **Reference:** Article 15 – Cybersecurity requirements for AI systems. - **Data Transparency** - **Reference:** Article 13 – Documentation requirements for high-risk systems. - **Consent** - **Reference:** Article 52 – Consent and communication obligations for users. --- ### **AI-Driven Learning Tools** - **Content Accuracy** - **Reference:** Article 10 – Dataset quality assurance. - **Personalization** - **Reference:** Article 14 – Aligning personalization with ethical oversight. - **Feedback Mechanism** - **Reference:** Article 54 – Reporting and feedback mechanisms for AI systems. - **Inclusivity** - **Reference:** Article 10 – Diverse datasets to ensure inclusivity. - **Cultural Sensitivity** - **Reference:** Article 9 – Risk mitigation strategies, including cultural sensitivity. --- ### **Grading and Assessments** - **Fairness** - **Reference:** Article 7 – Prohibitions of certain AI practices (unfair grading systems). - **Explainability** - **Reference:** Article 14 – Human oversight to ensure explainability. - **Error Correction** - **Reference:** Article 56 – Error correction and liability mechanisms. - **No Sole Decision-Making** - **Reference:** Article 14 – Human oversight in decision-making processes. - **Accuracy Validation** - **Reference:** Article 10 – Continuous testing for dataset and model accuracy. --- ### **Student Engagement** - **Interaction Design** - **Reference:** Article 14 – Ensuring systems are designed for responsible usage. - **Feedback Personalization** - **Reference:** Article 9 – Personalization with fairness considerations. - **Privacy in Chatbots** - **Reference:** Article 52 – Transparency in AI communication tools. - **Monitoring Usage** - **Reference:** Article 15 – Usage monitoring and cybersecurity protocols. - **Emotional AI Limitations** - **Reference:** Article 5 – Prohibition of harmful or manipulative AI systems. --- ### **Training for Educators** - **AI Literacy** - **Reference:** Article 9 – Training and awareness programs for stakeholders. - **Bias Awareness** - **Reference:** Article 10 – Educator training on dataset biases. - **Decision Oversight** - **Reference:** Article 14 – Human decision-making training requirements. - **Ethical Use Training** - **Reference:** Article 9 – Awareness programs on ethical AI use. - **Policy Awareness** - **Reference:** Article 23 – Internal policies for AI governance in organizations. --- ### **Procurement and Deployment** - **Supplier Compliance** - **Reference:** Article 24 – Supply chain management obligations. - **Documentation** - **Reference:** Article 13 – Comprehensive documentation of AI system operations. - **Risk Assessment** - **Reference:** Article 9 – Risk management plan for deployment. - **Third-Party Audits** - **Reference:** Article 20 – Conformity assessments by third parties. - **Regular Updates** - **Reference:** Article 13 – Documentation to reflect system updates. --- ### **Special Needs and Accessibility** - **Accessibility Features** - **Reference:** Article 14 – Inclusion of human oversight for accessibility. - **Support for Disabilities** - **Reference:** Annex III – High-risk systems, including those designed for disabilities. - **Language Support** - **Reference:** Article 10 – Dataset diversity, including multilingual data. - **Customizable Interfaces** - **Reference:** Article 9 – Designing customizable features for inclusivity. - **Assistive AI Validation** - **Reference:** Article 20 – Validation and testing for assistive technologies. --- ### **Long-Term Impact** - **Future-Proofing** - **Reference:** Article 13 – Regular updates and adaptable documentation. - **Sustainability** - **Reference:** Article 9 – Environmental considerations in AI usage. - **Cost-Benefit Analysis** - **Reference:** Article 20 – Economic assessment during conformity checks. - **Scalability** - **Reference:** Article 13 – Design requirements for scalability. - **Continuous Improvement** - **Reference:** Article 54 – Reporting and adapting AI systems. --- ### **Stakeholder Engagement** - **Parental Involvement** - **Reference:** Article 52 – Transparency obligations to inform all stakeholders. - **Student Participation** - **Reference:** Article 54 – Mechanisms for user feedback. - **Community Outreach** - **Reference:** Article 23 – Institutional governance including community input. - **Cross-Institution Collaboration** - **Reference:** Article 20 – Shared practices through audits and testing. - **Regulatory Liaison** - **Reference:** Article 62 – Coordination with regulatory authorities. # Article 6 – Classification of AI Systems
#### **1. Four Risk Levels** AI systems are classified into **four categories** based on their risk potential:
- **Prohibited AI Systems:** AI practices that pose unacceptable risks and are banned. - Examples: Subliminal manipulation, exploitation of vulnerabilities, and social scoring systems by public authorities. - **High-Risk AI Systems:** Systems that significantly impact individuals' safety or fundamental rights. - Examples: AI used in biometric identification, critical infrastructure, education, employment, credit scoring, or healthcare. - **Limited-Risk AI Systems:** Systems requiring transparency obligations but not as tightly regulated as high-risk systems. - Examples: Chatbots, recommendation systems, and virtual assistants. - **Minimal-Risk AI Systems:** Systems with negligible risk, which are largely unregulated. - Examples: Entertainment AI, spam filters, and AI-powered games. ---
#### **2. Criteria for Classification** The classification process considers:
- **Sector of Application:** The domain where the AI is deployed (e.g., education, healthcare). - **Impact on Rights and Safety:** How the AI affects individuals' privacy, safety, or fundamental rights. - **Severity of Harm:** The potential damage caused by incorrect or biased outcomes. - **Autonomy of AI Decision-Making:** The level of human involvement or oversight in the AI's decisions. ---
#### **3. High-Risk Systems in Education** Educational AI systems are explicitly listed under **Annex III** of the EU AI Act as high-risk if they:
- Determine **student access to education** (e.g., AI used in admissions). - Influence **learning outcomes** (e.g., AI-powered grading systems). - Assess **skills or competencies** that significantly affect career prospects.
These systems must comply with strict regulations, including documentation, risk management, and transparency requirements.
---
#### **4. Obligations for High-Risk Systems** For AI systems classified as high-risk, the following obligations apply:
- **Conformity Assessments:** Systems must pass pre-deployment evaluations for compliance. - **Risk Management:** Continuous risk assessment throughout the system's lifecycle. - **Data Requirements:** High-quality, representative, and bias-free datasets. - **Human Oversight:** Mechanisms to ensure human intervention when needed. - **Monitoring and Reporting:** Continuous monitoring of performance and reporting of incidents. ---
### **Implications for Stakeholders** #### **AI Developers**
- Must evaluate whether their system falls under high-risk categories. - Implement safeguards like robust testing and documentation.
#### **Educational Institutions**
- Ensure AI tools for admissions, grading, or skill assessment meet high-risk criteria. - Conduct regular audits to verify compliance with the EU AI Act.
#### **Regulators**
- Monitor the deployment of AI systems in high-risk areas, especially those influencing fundamental rights. - Enforce penalties for non-compliance. ---