CESI at SEISMEC to share trade union views on human-centric ICT at work

In a discussion with the consortium members of the EU co-funded SEISMEC project on the development of worker-focused ICT and AI tools in firms and companies, CESI laid out key concerns and issues that need to be addressed to secure employees' trust and acceptance of new digital tools at work.

CESI stressed in particular: 

• ICT and AI bring automation and can boost productivity and growth - but also have the potential to bring job disruptions and job displacements. This necessitates an inclusive and comprehensive approach to re-skilling and up-skilling workers to new job roles and functions that ICT/AI create. It also necessitates strong social security nets for those workers that, for various reasons, cannot be retrained and will drop out of work.
• Where ICT/AI tools are deployed for workforce organisation - such as for the tracking of working time or for hiring practices, job evaluation and promotion exercises and/or dismissals - humans must stay in control of ICT/AI tools and be able to challenge them. Especially in these cases, ICT/IT tools should improve and augment human decision making, but not replace it.
• To alleviate employees' mistrust and increase their acceptance of new ICT/AI tools at work, ICT/AI tools must be fed with unbiased data and employers must communicate and explain proactively and in understandable ways why and how they are implemented. Transparency and accountability are key.
• Deployed ICT and AI tools must constantly and continuously be reviewed in audits to identify and correct any biases in algorithms. Clear protocols for addressing any issues found during audits is crucial, and there must be independent mechanisms for employees to raise concerns about the functioning ICT and AI tools at work.
• Information and consultation frameworks and social dialogue with social partner agreements provide relevant fora to discuss new ICT/AI tools and their development from the outset together with workers and their representatives, ideally on the basis of ex-ante impact assessments.