ILO paper explores how international labour standards can guide ethical AI at work

The analysis finds substantial common ground around fairness, human dignity, well-being and the prevention of harm, establishing a basis for the two frameworks to inform each other as AI becomes more widely used at work.

International labour standards can help turn broad principles of artificial intelligence ethics into practical protections at work, according to a new International Labour Organization (ILO) working paper examining how existing rights can guide the fair and responsible use of AI.

Written by ILO researchers Jordi Agustí-Panareda and Daniel Samaan, AI Ethics and the ILO's international labour standards system: exploring synergies examines the relationship between widely recognised AI ethics principles and the values, principles and rights underpinning the organisation's labour standards.

Many AI ethics guidelines introduced in recent years remain broad and voluntary, providing limited direction for specific workplace situations, and the researchers identify international labour standards as an important reference for addressing decisions that affect workers' rights, dignity and working conditions.

Labour standards bring collective rights into AI discussions

The researchers compare five recognised AI ethics principles, justice, beneficence, non-maleficence, autonomy and explicability, with the ethical foundations of the ILO's standards, examining questions of fairness, promoting well-being, preventing harm, respecting individual choice and making AI understandable.

The analysis finds substantial common ground around fairness, human dignity, well-being and the prevention of harm, establishing a basis for the two frameworks to inform each other as AI becomes more widely used at work.

It also identifies differences in emphasis, with AI ethics frameworks often concentrating on individual autonomy, privacy and transparency, and the ILO's framework giving additional attention to collective rights, including freedom of association, collective bargaining and social dialogue.

These collective dimensions are frequently absent from mainstream AI ethics discussions, despite their importance in giving workers a voice in decisions about how technology is introduced and governed in their workplaces.

Hiring, surveillance and worker data raise practical questions

The paper examines how the collection and use of workers' personal data, algorithmic management, workplace surveillance and AI-assisted employment decisions raise questions that need renewed attention when applying international labour standards.

When employers use AI to screen job applicants or assess employee performance, understanding how a decision was reached and what data informed it becomes essential to accountability and a fair process consistent with those standards.

Such uses bring ethical principles into everyday employment decisions, where questions about transparency and the prevention of harm have direct consequences for people seeking jobs or having their work evaluated.

The authors suggest that established labour rights can help shape workplace AI governance, and that emerging AI ethics principles can draw attention to concerns that inform the application of labour standards as technology develops.

Governments, employers and workers all have a role

The study highlights the contribution that the ILO's international labour standards system and tripartite structure can make to AI governance, bringing governments, employers and workers into discussions about technological change.

This structure provides a way to consider AI's effects on working conditions, collective representation and social justice alongside its technical features and consequences for individuals, giving workplace governance a broader foundation.

The authors call for further research and dialogue on how AI ethics and international labour standards can reinforce each other, supporting an approach to technological change that places people and their rights at the centre of decisions about work.