PCS AI campaign to ensure workplace justice

AI will continue to have a big impact on the world of work as well as in wider society. The NEC this week agreed to develop a bargaining plan centred on our AI model agreement and negotiations with the Cabinet Office. PCS wants to ensure that the introduction and use of AI is to ‘improve, not dehumanise’ work.

Recent reports suggest AI could bring real efficiencies to civil service departments. It is essential that there is transparency about how AI systems work and decisions about how they are used. Public trust demands this, as would workers overseeing AI-driven systems. Transparency will help ensure that AI is used fairly and potentially for good and with high levels of accountability.

We believe is important that any efficiencies made by AI are used to improve services rather than simply cutting the cost of existing service levels through mass redundancy, as previous cabinet secretary Oliver Dowden stated.

The rights of workers and unions should be protected and PCS is supporting the TUC’s draft bill which aims to do just that. We are calling for consultation and equality impact assessments to be carried out.

The bill raises concerns and proposes measures to ensure transparency as well as dangers around safeguarding, discrimination and regulation. And where AI is used for recruitment, HR, work allocation or disciplinaries, there has to be secure and fair-minded decision-making processes.

AI model agreement

Our model agreement says that new AI systems should only be introduced following consultation and agreement with the union. And all existing AI systems should be disclosed to the union. 

It is explicit that AI must not lead to job losses, although we recognise it will impact on job roles, and that AI is used to enhance jobs and not to degrade them and therefore prevent humans being reduced to ‘machine minders’.  

It goes on to say no AI system will be introduced, or if in existence, used, if in part or whole it is designed to make any decision that affects humans, in particular to select staff for recruitment, promotion, transfer to another post or dismissal.

It should not remove existing decision making from humans or automate decision making. Nor should it be used to allocate, monitor or evaluate the work of humans or undertake any form of surveillance of humans.

The agreement says that AI systems should not affect the pay, terms and conditions and status of staff unless such changes are agreed by both parties and that there will be no downgrading of roles as a result of AI.

Lastly that productivity gains will be made through the introduction and use of AI, but these gains should be shared with workers in the form of reduced hours and increased pay.