Employment Rights Bill

Baroness Bennett of Manor Castle Excerpts

Committee stage
Thursday 5th June 2025 Lords Chamber

[Lord Holmes of Richmond (Con)]

My Lords, it is a pleasure to follow my friend the noble Lord, Lord Clement-Jones. In doing so, I declare my technology interests as set out in the register. It is a pleasure to follow him because this has always been his “WAIRIA” of expertise—bear with me. I will speak to my Amendments 289 to 298 and 314 to 316, but before doing so, I give full-throated support to everything the noble Lord said and his amendments. We are very much on the same page.

There is a strange situation with government at the moment when it comes to AI. That is not specific to employment rights but across the piece. We have been subject to it for the past year. We are told consistently that the Government will not be bringing forward cross-sector AI legislation. That position is to be defended if it is taken—the Government have decided on a domain-specific AI approach. But the difficulty with that is that whenever we have had domain-specific legislation coming through your Lordships’ House—be it product regulation, data or any of the Bills that I, my friend the noble Lord, Lord Clement-Jones, and others, have worked on—we have been told that those are not the Bills where AI is to be considered. In only a slightly reductive way, we currently have a situation, to be clear, where the Government are saying they are not bringing forward cross-sector AI legislation and specific Bills are largely—not exclusively—not the place to incorporate AI issues.

The amendments that noble Lord, Lord Clement-Jones, and I set out in this group are key to one of the most important sectors—it is broader than a sector, and such an important aspect of our lives. It is how we are employed, what that employment looks and feels like, and how it is experienced by all of us. These amendments do not seek to address issues that will occur next year, next month or even tomorrow. AI is impacting workers right now, oftentimes without them even knowing that it is in the mix.

My first amendment seeks to suggest that the principles that have variously appeared in White Papers and other reports are put on a statutory basis in the Bill. We give ourselves the best opportunity to optimise with AI if we take a principles-based, outcomes-focused and input-understood approach. Similarly, I set out in Amendment 290 that all employers and organisations that develop, deploy or use AI should have an AI responsible officer. For this, do not think burdensome, bureaucratic or overcompliance. Because of the proportionality principle, it simply means that there is an obligation on those employers to report on their use of AI in the workplace. It can be well understood through reporting obligations such as those set out in the Companies Act, which employers will be very familiar with at this stage.

My amendments then move to questions of use. What happens where IP or copyrighted material is being used in the workplace? There needs to be labelling so that everybody is clear on, and there is transparency about, what is going on. What about the use of workers’ data? This is an incredibly rich resource that should not in any sense be served up or sold off to the highest bidder. The use of AI in the workplace should be clear and transparent, and workers should have an opt-in, not an opt-out, responsibility, as set out in the amendments.

Then, as the noble Lord, Lord Clement-Jones, has touched on, there is the question of automated decisions. It is clear that workers not only have to be aware that ADM is being used—and have the right to opt out—but also need the right to a human explanation of what is happening in those situations. If we are to optimise things with these technologies, concepts such as “human in the loop” and “human over the loop” must be understood. Safeguards need to be in place, not least where ADM is used, and this could form part of the data protection impact assessment that employers have to undertake.

Then there is the question of regulators. Employment and recruitment currently find themselves wide open to the use of AI. An individual may find themselves not getting shortlisted, not getting hired and not even knowing that the reasoning behind that was algorithmic processing rather than human judgment and human reasoning. It is critical to consider the right approach to fill that regulator gap. Would a specific employment and recruitment regulator do the job? My view—and I think there is evidence to support this—would again be that we could have a cross-sector AI authority. Again, do not think of a bureaucratic and burdensome AI regulator; instead, think of a nimble, agile, adaptive and, crucially, horizontally focused AI regulator, not only in the area of employment rights but across the whole of our economy and society. It would deliver that clarity, consistency and certainty that we all need wherever we come across AI in our working, professional and private lives.

It is so significant that, in Amendment 315, I believe there should be a commission on AI in the workplace. Mindful of comments from Monday, I am certainly no fan of setting up a commission to delay or kick issues into the long grass. But perhaps by using the technology to solve some of the issues that are created by the technology, we could have a reimagined approach to commissions and consultations.

Finally, I come to Amendment 316 and the algorithmic allocation of work. This is already happening, and it has already been in front of the courts. It is clearly an issue and one that needs to be fully understood. The Government need to state clearly their position on this most significant of matters. I look forward to other speakers and to the Minister’s response.

[Baroness Bennett of Manor Castle (GP)]

My Lords, it is a pleasure to follow two of the House’s acknowledged experts in this area of the impact of AI. I will speak to my own Amendment 323B and also note that I attach my name to Amendments 294 and 298 in the name of the noble Lord, Lord Holmes.

My Amendment 323B is quite a modest step. It calls for a review to be published within 12 months. In saying that, I thank the Ministers for having a meeting prior to the discussion of these amendments, which I very much appreciated. But I think the time for talk is over; the time for action is now. Twelve months is still too short, but it seemed the best timeframe I could reasonably give for this call for a review of the electronic monitoring of workers in the workplace. This picks up some points made by the noble Lord, Lord Holmes. It also crucially points to the need to look around the world and see what else is happening and what we can learn from what has happened in other places. The companies selling these systems are global giant multinational companies. The companies deploying these systems are giant multinational companies in many cases. It is important that, rather than trying to pick this off ourselves, we look around the world and say that we want to be leaders in creating a different kind of model of how workers can be protected.