The growing use of algorithmic management tools has led to profound changes in labour relations. These systems, based on artificial intelligence (AI), are deployed not only to increase productivity within companies and institutions but also to strengthen managerial power through sophisticated technological instruments for the digital monitoring of work activities.

Algorithmic management systems collect personal data and, in some cases, can analyse workers’ behaviour and emotions, affecting labour relations, by handling hiring processes, the distribution of tasks, the monitoring and control of workers, performance assessment and decisions on promotions and dismissals, be they individual or collective.

The lack of transparency regarding algorithmic programming criteria applied in the world of work is steadily giving rise to a growing imbalance between workers’ level of knowledge (about the handling of the digital data collected) and that of the decision makers in the work environment (who use these tools to obtain consolidated information and analyses on matters that affect and control work relations).

Beyond the implications in terms of respect for the principles enshrined in national and international regulatory instruments, such as transparency and explainability, tackling algorithmic discrimination (due to biases of various kinds) requires collective, coordinated and strategic action by the labour movement.

Algorithmic discrimination

Examples of rights violations through the application of AI tools abound. Initiatives to tackle these practices can also be identified and trade unions have a crucial role to play in this respect.

In a landmark case of algorithmic discrimination in Italy, the workers invoked the Council of Europe’s Data Protection Convention 108+ and the EU’s General Data Protection Regulation (GDPR) to identify a food delivery app’s algorithm as discriminatory.

The algorithm managed workers’ shifts according to their “reliability” – as measured by shift assiduity – without taking into account the reasons for missing a particular shift, such as being ill.

In the United States, the American Civil Liberties Union filed a complaint with the US Equal Employment Opportunity Commission against Meta (Facebook), noting that the company allowed gender-targeted advertisements, targeting only male users through the use of recommendation algorithms.

Another example of algorithmic management with discriminatory programming was identified in the Predictive Hiring application, which seeks to avoid potential future costs for companies through predictions based on workers’ data. This algorithm rules out employees that are more likely to ask for wage increases or support unionisation, by transferring and analysing a range of data compiled on workers through online platforms.

Handbook for collective bargaining on algorithms

The handbook for collective bargaining on algorithms produced by Spain’s UGT union confederation, accessed by Equal Times, is a concrete example of how trade unions can prepare and mobilise to address the challenges posed by AI in the workplace. It offers guidance and practical tools for trade unions in negotiating collective agreements that specifically address the impact of AI on labour relations and workers’ rights protection.

These agreements recognise the right of workers not to be the object of decisions based solely on automated variables, as well as the right to non-discrimination in relation to decisions and processes based solely on algorithms.

One of the cases highlighted in the handbook is the XXIX Collective Bargaining Agreement signed at El Norte de Castilla, S.A. (2021-2023), which includes specific provisions regarding the introduction of new technologies and their impact on working conditions.

The agreement stipulates that “the company must inform the Works Council in advance of any plan to introduce new technologies that may modify the working conditions of the various professionals working for El Norte de Castilla, S.A.”. It also states that “the introduction of new technology will not lead to a reduction in the number of permanent employees”.

Another significant example is the collective bargaining agreement at Air Nostrum Training Operations, which specifically addresses the use of algorithms in labour relations, establishing that the Works Council must be informed in advance of any plans the company has to use decision-making software or algorithms in labour processes of any kind.

Similarly, the National Collective Bargaining Agreement for companies and workers in the perfume and related industries and the 24th Collective Bargaining Agreement for the banking sector contain provisions on workers’ digital rights in relation to AI.

Collective bargaining on algorithms in the US: from the NBA to electricity workers

In the United States, a study on Union Collective Bargaining Agreement Strategies in Response to Technology, conducted by the Center for Labor Research and Education at the University of California, Berkeley, presents trade union responses to technological change in the country. The paper describes how unions are leveraging their collective actions and bargaining agreements to address the digital and technological transition.

The study highlights the provisions within collective bargaining agreements that seek to establish rights and responsibilities in the adoption of digital tools and looks at clauses designed to mitigate the effects of technological change in the workplace.

Of the collective agreements examined, one that stands out is the CBA covering players in the United States National Basketball Association (NBA).

Signed in 2017 and renewed until 2030, this agreement sets limits on the use of information obtained through monitoring technologies in the evaluation of players, particularly for the negotiation of salaries. The NBA contract guarantees the right of unions and employees to take part and collaborate in decision-making on monitoring technologies. It also establishes a joint advisory committee to review monitoring practices and technologies and the use of sensor-derived data.

The agreement includes a provision that allows the union to appoint experts who can approve devices and establish the cybersecurity standards to be associated with them. It also stipulates that the data collected by sensors can only be used for sporting ends and not for economic or contract bargaining purposes.

Following the signing of this CBA for the NBA in 2017, other unions have included algorithms in collective bargaining. In 2019, for instance, United Auto Workers (in the US) negotiated a contract with General Motors that included provisions on the use of artificial intelligence and algorithms in the workplace. The contract protects workers from disciplinary action or dismissal as a result of decisions made solely by an algorithm.

Similarly, the International Brotherhood of Electrical Workers (IBEW) signed agreements with various utility companies to ensure that the use of algorithms does not result in job losses for its members.

These examples highlight the importance of including specific clauses on algorithms and AI in collective agreements to ensure the protection of labour rights and the proper management of technology in the workplace.

They also reflect an increased awareness of the ethical and labour challenges associated with the use of AI in the workplace, and the need to address these challenges through social dialogue and collective bargaining.

Agreement between AFL-CIO and Microsoft

In the United States, at the end of 2023, the American Federation of Labor and Congress of Industrial Organizations (AFL-CIO) and Microsoft Corp. announced a partnership to discuss how artificial intelligence should look at the needs of workers and include their voices in its development and implementation.

This is the first agreement focused on AI between a labour organisation and a technology company. Its goals are threefold: to share detailed information with union leaders and workers on AI technology trends; to incorporate workers’ perspectives and expertise in the development of AI technology; and to help shape public policy that support the technology skills and needs of frontline workers.

Key provisions of the agreement include a commitment to respect the right of employees to form or join unions, to develop positive and cooperative labour-management relationships, to negotiate collective bargaining agreements that will support workers in an era of rapid technological change, and to expand the role of workers in creating worker-centred design, job training and trustworthy AI practices.

Another pillar of the partnership between AFL-CIO and Microsoft is the establishment of feedback mechanisms in which union representatives share knowledge and concerns directly with the developers of the AI products, creating opportunities for the co-design and development of worker-centred technology ahead of its implementation.

The work plan of the agreement between the AFL-CIO and Microsoft sets out a roadmap for AI education for workers and students starting in 2024. It also includes the development of joint policies to increase training opportunities in non-traditional technology occupations.

TUCA’s trade union strategy on AI

To enhance debate on trade unionism in the context of digitalisation, the Trade Union Confederation of the Americas (TUCA) has published a series of bulletins and reports through its Labour Observatory of the Americas (OLA).

The papers address issues such as the use of artificial intelligence (AI) and algorithmic labour management, providing the trade union movement of the Americas with analytical and strategic tools to address these challenges. The TUCA has also created spaces for reflection and formulation within working groups such as the Hemispheric Development and Integration Working Group (GTDIH), which holds activities on the implications of AI development and the challenges this technology poses for the working class.

Such efforts are geared towards gaining a better understanding of the impact of digitalisation on the workplace and designing strategies to ensure that the implementation of new technologies benefits all workers, promoting equity, social justice and sustainable development in the workplace.

The challenge of globalising trade union action on the impacts of AI

In Europe, there are already collective agreements that regulate the use of technology, both with regard to the surveillance of workers and the distribution of tasks. The trade union centres, based on the AI Act of the European Union, foresee the potential of collective bargaining as a micro-regulatory instrument for issues such as data protection, privacy and the right to information on the use of algorithmic tools and their programming criteria.

In the United States, the strategic action being taken by trade union centres aims to influence both the usability and the origin of the design, testing and application of AI tools through coordinated action with businesses, academia and political actors.

In other continents, academia is playing an essential role in developing the inputs required for the incorporation of the key concepts and the trade union training activities aimed at developing action strategies for trade union centres.

The initiatives undertaken by the TUCA are an example of the effort to coordinate the formulation of positions and collective bargaining strategies on algorithms in the Americas.

The cooperation actions of the International Trade Union Confederation are crucial to extending the initiatives and good practices developed in Europe, the Americas and the United States to other parts of the world.

Globalising trade union activism to tackle the impacts of AI on the world of work is central to ensuring a just and sustainable future of work. As AI continues to transform our economies and societies, trade unions have a vital role to play in defending workers’ rights and promoting the ethical governance of technology.

Role of the trade union movement in the programming and implementation of AI systems

The unregulated application of AI tools in the world of work and beyond offers the workers’ movement a unique opportunity to boost its revitalisation efforts. Through communication initiatives and the promotion of training activities on digitalisation issues, trade unions have the potential to be at the forefront of debates on artificial intelligence, its impact on the world of work and social protection for workers.

Strategic action for the inclusion of specific clauses on algorithms and AI in collective agreements is a key factor in ensuring effective pathways towards labour rights protection and the proper management of technology in the workplace.

In this respect, the emergence of partnerships between the labour movement and actors such as academia and companies developing AI tools is key to adapting trade unions to the digital age.

Alongside the development of inter-institutional partnerships, collective bargaining agreements are undergoing a process of substantive adaptation, with the alteration of their contents to regulate the use of AI tools and algorithms in the workplace, recognising the importance of protecting workers’ rights and promoting a fair and equitable labour relationship for the application of ethical principles within the framework of responsible AI.