AI: Multifaceted yet manageable challenge for workers and trade unions

In our ‘think like a union’ roundtables, we tried to explain some of the main challenges that unions face with the introduction of AI as well as discussing its impacts, and how these relate to the demands of unions to minimise negative impacts on working conditions and labour rights, while responding to questions and perspectives from the participants. 

We started by summarising the role of unions as representative organisations for workers, listening to them, and taking up the problems they have with employers, or – depending on local structures of social dialogue – also with government. Unions therefore have to operate at multiple levels and respond to their own members, as well as negotiating with employers and discussing with government authorities, which adds complexity to their tasks. Then we discussed the direct and indirect impacts that digital technologies in general can have on workers. We reviewed the various legal categories under which workers’ issues with technology are traditionally discussed; occupational safety and health, discrimination, social protections, privacy, wages and contractual issues and so on, and how these are largely unknown among those that develop new technologies. This means that AI developers tend not to consider these issues when thinking of the impacts.

In our discussions we highlighted some of the problems that digitalisation, and associated tools like AI, bring to workers. One example discussed was of an AI system that had negative health and safety impacts on workers. Workers approached management about the issues, but the managers explained that they did not set the parameters used by the system. After some time the union was told to contact another company based abroad, and when this company was eventually contacted, the union was told that the parameters were actually set in a third country. Naturally, this created a substantial delay to the process of resolving the problems. This case highlights the ways in which AI, as a work tool, can complicate employment relationships when management are no longer in direct control of all the ways in which work is being governed.

Another issue highlighted during the roundtable was the way in which traditional regulation tends to see new technology as an issue within an employer’s discretion, while work rules are subject to negotiation. But with AI, a work tool effectively transforms and governs work rules, and does so in ways that are not transparent to workers, and which can be changed by the click of a mouse. So unless AI is opened up to negotiation, it can remove key areas from the realm of collective bargaining. 

The participants imagined themselves as workers in particular sectors and then considered the technology-related problems that these workers face. We discussed the sources of these problems, including the lack of any say in the design of the technologies, the opacity of algorithmic decision-making, the potential for discriminatory decisions, and issues relating to isolation, surveillance and fatigue management, the changes to performance management and the double-edged nature of many technological solutions which may increase security in one instance, but lead to increased stress or discriminatory outcomes in another. 

There is sometimes a perception that unions are only interested in saying ‘no’ to technology, but I hope that the discussions in the roundtables helped participants to better understand the many dimensions of the tech problem for workers and their trade unions. We highlighted that measures – such as those called for in the ITF Technology and Decent Work Charter – may help to resolve these problems and enable unions and employers to work effectively together to ensure frictionless tech deployments that sustain innovation and preserve and protect the rights of all.