Fairness
AI adoption should not systematically disadvantage particular groups of workers or learners. Organisations should examine both unequal access to AI tools and discriminatory outcomes from their use.
Work & employment IWIB4AI session Needs review
SPOT Nordic Annual Meeting 2025
Using AI in work and education while protecting fairness, human agency and opportunities to learn.
Draft – legal and case law references not yet verified
Summary
This use case concerns the ethical integration of AI into work and education, including its influence on tasks, skills and learning. AI can support productivity and learning, but its benefits depend on how organisations address unequal access, changing skill requirements and human control. The central challenge is to support workers and learners rather than simply automate processes or shift responsibility to technology.
The Problem
How can employers and educators adopt AI without weakening human judgment, widening inequalities or limiting opportunities? The tension is between efficiency gains and the need to preserve fair treatment, meaningful learning and accountability for consequential decisions.
Ethical Grounds
Autonomy
Workers and learners should retain meaningful scope to exercise judgment rather than become dependent on AI recommendations. Adoption should support skill development and informed choices.
Transparency
People should understand when AI influences their work, learning or assessment and what its limitations are. Clear explanations help them question outputs and recognise when human review is needed.
Human oversight
Consequential decisions affecting employment or educational opportunities need competent human scrutiny. Oversight should include the authority and practical ability to challenge or override AI outputs.
Legal Compliance
| Framework | Provision | Relevance |
|---|---|---|
| EU AI Act | Art. 6(2)–(3); Annex III, points 3–4 | Certain education and employment systems, including specified assessment, recruitment and worker-management applications, qualify as high-risk, subject to the Act's classification conditions and exceptions. |
| EU AI Act | Art. 4 | Providers and deployers must take measures to ensure sufficient AI literacy among staff and others operating AI systems on their behalf, taking account of their knowledge and the context of use. |
| GDPR | Art. 5(1)(a)–(c) | Processing workers' or learners' personal data must be lawful, fair and transparent, limited to specified purposes and restricted to necessary data. |
| GDPR | Art. 22 | Solely automated decisions producing legal or similarly significant effects on workers or learners are generally prohibited unless an applicable exception and required safeguards apply. |
Sources
Court & Regulator Decisions
FILCAMS CGIL et al. v Deliveroo Italia Needs review
Rankings that ignored reasons for absence illustrate how efficiency-focused workplace algorithms can discriminate without explicitly using protected characteristics.
IMY – Skellefteå municipality facial recognition in school (opens in a new tab)
Unlawful biometric attendance monitoring illustrates how educational AI can undermine learner agency when institutions mistake participation or consent for genuine choice.
Drivers v Uber (algorithmic dismissal) (opens in a new tab)
Automated worker deactivation directly informs safeguards for consequential workplace decisions, including transparency and protection against solely automated decision-making.
Key Takeaways
- Assess each AI application separately: routine assistance and consequential decision-making carry different risks.
- Build role-specific AI literacy for staff, workers and educators.
- Test for unequal impacts and provide meaningful ways to challenge consequential decisions.
- Design AI adoption to preserve human judgment, skill development and access to learning.
Related Use Cases
From the Session
SPOT Nordic Annual Meeting 2025