Artificial intelligence can expand access to education, support educators and make knowledge available across languages and borders. It can also reproduce bias, weaken privacy, obscure accountability and create new risks for learners and institutions.
At ICARUS, responsible AI is therefore not treated as a technical feature. It is a governance responsibility.
Our Responsible and Ethical AI Policy establishes the principles, safeguards and accountability structures that guide how we develop, select, integrate and use artificial intelligence across our educational services and internal operations.
Our objective is not simply to make AI more capable. It is to ensure that AI strengthens human capability, protects individual rights and contributes to more equitable access to education.
Our responsible AI principles
ICARUS applies eight principles throughout the AI lifecycle:
Human agency and oversight
AI should support human judgment, not displace it. Meaningful human oversight is required whenever AI may affect learning outcomes, access to opportunities, employment, privacy, safety or other significant interests.
Educational benefit and proportionality
Every AI application must have a legitimate and clearly defined purpose. The expected educational or operational benefit must be proportionate to its foreseeable risks.
Fairness and inclusion
We seek to prevent discriminatory outcomes and reduce bias across languages, cultures, abilities, socioeconomic circumstances and levels of digital access.
Privacy and data protection
Personal data must be used lawfully, fairly and transparently. ICARUS applies data minimisation, purpose limitation, appropriate security, retention controls and safeguards for automated decision-making where applicable.
Transparency and explainability
People should be informed when they are interacting with AI or receiving materially AI-generated or AI-supported content. Explanations should be appropriate to the system, its context and the people affected.
Safety, security and reliability
AI systems must be evaluated for accuracy, robustness, cybersecurity, harmful outputs and foreseeable misuse before deployment and throughout their operational lifecycle.
Accountability and remedy
Every material AI application must have a named human owner. Where ICARUS causes or contributes to an adverse impact, we are committed to taking proportionate corrective action and enabling appropriate remedy.
Environmental responsibility
We consider the energy and resource implications of AI technologies and favour solutions proportionate to the educational or operational purpose.
How ICARUS puts these principles into practice
Our governance approach includes:
- An inventory of material AI systems and applications
- A named accountable owner for each material AI use case
- Proportionate AI, educational, privacy, security and human-rights impact assessments
- Human review for higher-risk or consequential uses
- Testing for accuracy, bias, accessibility, robustness and harmful outcomes
- Assessment of AI suppliers, data practices and contractual safeguards
- Clear information for learners, educators and other affected stakeholders
- Channels for questions, concerns, correction and challenge
- Monitoring of performance, incidents, complaints and material system changes
- Responsible-AI and AI-literacy training for relevant personnel and contractors
ICARUS will not deploy an AI system where material risks to human rights, safety, privacy, educational integrity or equal treatment cannot be reduced to an acceptable level.
Special responsibility in education
Education requires a higher standard of care.
AI-generated educational material may sound authoritative even when it is incomplete, biased or inaccurate. ICARUS therefore requires proportionate review by suitably qualified people whenever content could affect educational accuracy, learner safety or academic outcomes.
Our approach is designed to preserve:
- The essential role of educators
- Learner agency and critical thinking
- Academic and assessment integrity
- Accessibility and linguistic inclusion
- Meaningful feedback and human interaction
- The right to question or challenge consequential outcomes
AI should expand what educators and learners can achieve together. It should not reduce education to automated content delivery or place unreviewed technology between a learner and their rights.
Alignment with European and international frameworks
The ICARUS policy is designed to support compliance with applicable law and alignment with authoritative international frameworks.
European Union Artificial Intelligence Act
ICARUS follows the risk-based approach established by Regulation (EU) 2024/1689—the EU Artificial Intelligence Act.
Depending on ICARUS’s role and the nature of the application, this includes:
- Identifying prohibited and potentially high-risk AI practices
- Maintaining appropriate documentation
- Supporting AI literacy
- Providing transparency and human oversight
- Addressing accuracy, robustness and cybersecurity
- Monitoring material AI systems after deployment
European data-protection requirements
Where artificial intelligence involves personal data, ICARUS applies the principles of the General Data Protection Regulation and other applicable data-protection requirements.
This includes lawful and transparent processing, data minimisation, purpose limitation, security, individual rights and safeguards relating to automated decision-making.
UNESCO Recommendation on the Ethics of Artificial Intelligence
Our approach reflects the UNESCO Recommendation on the Ethics of Artificial Intelligence, adopted by UNESCO’s 193 Member States.
The Recommendation places human rights and human dignity at the centre of AI governance. Its principles include proportionality, safety, privacy, accountability, transparency, human oversight, sustainability, AI literacy, fairness and non-discrimination.
United Nations Guiding Principles on Business and Human Rights
ICARUS applies a human-rights due-diligence approach to material AI systems. This includes identifying potential impacts, taking action to prevent or mitigate harm, monitoring effectiveness and enabling or cooperating in remedy where appropriate.
United Nations Global Compact
As a participant in the UN Global Compact, ICARUS integrates its Ten Principles—covering human rights, labour, environment and anti-corruption—into the governance, procurement and use of artificial intelligence.
United Nations Global Digital Compact
ICARUS also supports the objectives of the UN Global Digital Compact: an inclusive, open, safe, secure and human-centred digital future that protects human rights, improves digital inclusion and strengthens the capacity of people and institutions to use AI responsibly.
Governance and accountability
The ICARUS Board or highest governance body oversees the policy and material AI risks.
Operational responsibility is shared across leadership:
- The Chief Executive Officer owns implementation and accountability.
- The Chief Technology Officer oversees architecture, security, testing, monitoring and supplier controls.
- The Chief Learning Officer oversees pedagogy, educational quality, accessibility, educator guidance and learner impact.
- Each material AI system has a designated use-case owner responsible for documentation, safeguards and periodic reassessment.
The policy is reviewed at least annually and following significant regulatory developments, material incidents or substantial changes to ICARUS AI systems.
From AI access to AI agency
Responsible AI in education cannot be achieved through compliance alone.
Learners and educators need the knowledge, confidence and institutional support to understand how AI affects them, evaluate its outputs and participate in decisions about its use.
For ICARUS, the long-term goal is therefore AI agency: enabling people to use artificial intelligence while retaining judgment, autonomy, cultural identity and meaningful control.
Important information sources:
- The EU AI Act and GDPR
- The UNESCO Recommendation on the Ethics of AI
- The UN Guiding Principles on Business and Human Rights
- The UN Global Compact Ten Principles
- The UN Global Digital Compact
- ICARUS AI Responsible & Ethical AI Policy
- Practical controls covering impact assessments, human oversight, vendor review, AI literacy, monitoring and remedy