HoC 85mm(Green).tif

Business, Energy and Industrial Strategy Committee 

Oral evidence: Post-pandemic economic growth: UK labour markets, HC 306

Tuesday 15 November 2022

Ordered by the House of Commons to be published on 15 November 2022.

Watch the meeting 

Members present: Darren Jones (Chair); Tonia Antoniazzi; Alan Brown; Mark Jenkinson; Andy McDonald; Charlotte Nichols; Mark Pawsey; Alexander Stafford.

Questions 65 - 116


I: Andrew Pakes, Deputy General Secretary and Research Director, Prospect; Brian Palmer, Head of Public Policy Europe, Amazon; Charlotte Holloway, Government Relations Director, Zoom; Neil Ross, Associate Director for Policy, techUK.

Examination of witnesses

Witnesses: Andrew Pakes, Brian Palmer, Charlotte Holloway and Neil Ross.

Q65            Chair: Welcome to this mornings session of the Business, Energy and Industrial Strategy Select Committee for our latest hearing in our UK labour markets inquiry, with a focus today on the role of artificial intelligence and technology in the workplace. I should say that we will be mentioning AI a lot today, but please do not limit your answers to a narrow definition of AI. We are interested in technology more generally.

On the first panel today, we are delighted to welcome Andrew Pakes, who is the deputy general secretary and research director at Prospect union; Brian Palmer, who is head of public policy in Europe for Amazon; Charlotte Holloway, the government relations director for Zoom; and Neil Ross, the associate director for policy at the trade body techUK. Good morning to all of you.

To kick us off, Neil Ross, there are two assumptions when we talk about these issues. The first is that big companies in the UK are pretty good at tech adoption and everybody else is pretty rubbish. The second is that, according to the Government recently, because of the shortage of people in the labour market, businesses should be investing in robots and technology to fill the skills gap in the workplace.

How true are those assumptions today? Has there been any recent change in those trends?

Neil Ross: Generally, those assumptions broadly hold true. If you look at the evidence, the highest productivity firms tend to be the largest. They tend to be based in London and the south-east. The lower productivity firms tend to be smaller and use technology less intensely.

This is a supertopical discussion. On Thursday, there are going to be two big themes that come out of the autumn statement when it comes to the labour market: participation and productivity. How do we get more people into the labour market, particularly after about 500,000 people have left over the course of the pandemic? How do we boost labour productivity, particularly at the lower end of the labour market?

Technology and AI have important roles to play in both. Participation technologies such as videoconferencing and hybrid working have an enormous role to play in getting more people into the workforce, particularly, those from disabled backgrounds or with mental health illnesses, who are largely underrepresented in the labour market and underemployed.

Technologies such as AI have a huge potential to boost productivity. Some recent research by Deloitte has highlighted that by 2035 AI could boost UK labour market productivity by 25%. Even if we got less than half of that, a 10% boost, that would do a massive amount to correct the issues we have around our debt and deficit.

AI systems that can boost productivity very significantly are really expensive and very difficult to use. You not only have to develop or buy the system itself; you also have to ensure you can train people to use it effectively. We have a massive digital skills shortage in this country. The Government estimate that the UK could lose up to £141 billion of GDP in total if we do not fix the digital skills shortage. Having such a big digital skills shortage puts enormous pressure on smaller companies in particular to retain key talent.

Over the last few weeks, we have been doing some surveys and focus groups with our tech sector members, so people who use these technologies a lot, and the smaller companies in particular are finding it very difficult to retain talent and key staff such as software engineers and developers. When you are a small company, losing two or three good members of staff can be really serious.

There is a big call-out to try to address the issues in the UK labour market. There are a number of ways of doing this. We need to improve the supply of high-skilled technical talent in the UK. We know from surveys by DCMS and others that people like data engineers and software engineers are not being churned out at the rate we need them from the education system, from universities etc. Therefore, we need to think about post-16 and post-18 retraining, whether that is through the apprenticeship levy or tax incentives for companies.

We also need to look internationally. We cannot fix this problem only through the domestic pipeline. That would take a long time. We also need to make sure we are bringing in the most talented people from around the world.

This is not just an economic issue. It is going to be a geopolitical and strategic issue. If the UK wants to lead on the foundational technologies of the future, artificial intelligence and quantum computing etc, we need to ensure the very best and brightest talent from around the world comes to the UK. That means looking at the immigration system as well. In effect, I am agreeing with your statement and then laying out some of the features that we see among our members.

Q66            Chair: Andrew Pakes, the Government said, “Where you cannot hire people to do the jobs, just buy some robots or technology”. From your perspective, does this trend mean anything for workplaces in reality? Was it just a statement that was not really reflective of what is going on?

Andrew Pakes: On the next panel, from my colleague from GMB, you will hear a lot more about some of those jobs that are most at risk or that are being augmented with digital technology. We are clearly seeing the development of technology, often in logistics and the just-in-time type economy, that is now beginning to stretch out into other parts of the job market. Post-pandemic, we have seen examples of things like shift allocation software and task allocation software coming to engineering and higher-skill STEM roles.

That does two things in this whole space. First, how is AI digital technology transforming the products and services that the UK delivers? Secondly—this is the one where we think there are more worrying tendencieshow is that technology being used to manage us? This is the socalled people analytics side. This is about products and services and then people management.

What we have seen during the pandemic is the acceleration of digital technology, which allows us to stay safe, connected and well. Without the likes of Zoom and their technology, Parliament would not have been to perform and many millions of people would not have been able to work. In that acceleration, we have also seen less time spent on the scrutiny of software or products that go into elements of surveillance and micromanagement.

A good example would be task allocation software, where you log in and it tells you how much time you have to do a task. We do not have clarity yet on how that data is then used in management processes to determine the speed of a job and whether you are a good or bad worker.

The first element is around transparency. How do we give people comfort? If the assumption of your previous question is that the future of this country is based on high-value and high-productivity goods, services and production, we would agree with that. If that is true, we also need a following assumption, which is that the labour market itself needs to be dynamic and help equip people for change.

Around three-quarters of our 2032 workforce are already in work. Lots of this debate is around young people, college and training, which is really important, but we also need a transition plan to enable those people already in work to see their future. The risk is that, as we have seen in other industries, when good high-quality jobs go in some parts of the country, they are often replaced by lower-quality, less respectful types of employment. Where do workers fit into that conversation?

Q67            Chair: Charlotte Holloway, there is this idea that big companies are good at tech and small companies are not so good. Presumably that has changed during the pandemic. Lots of them have been using Zoom, I assume.

Charlotte Holloway: Yes. Thank you for inviting us to the Committee today. It has been a really interesting journey for Zoom, seeing how businesses of all sizes and, indeed, the public sector and third sector have adopted hybrid and remote technologies. You are right: we saw adoption over that period, which showed that a decades worth of digital adoption and acceleration can happen in the space of two years. There has been a wholesale shift. The genie is out of the bottle for significant parts of the economy in utilising certain types of hybrid and remote technologies.

On the small business point, we actually found in a Zoom survey conducted with Boston Consulting Group last year that the UK was the highest adopter of hybrid and remote technologies of any significant developed country that we surveyed. The UK is very good at this stuff. We have a savvy consumer base. That translated in some ways to the way some of our SMEs adapted to technologies like Zoom during the pandemic.

Do we have further to go on the UKs digital skills base? Yes, we do, but the UK has really good foundations to be a global leader in hybrid and remote technologies. We think there are elements of the UK economy with certain industrial strengths—our knowledge economy, professional services and financial servicesthat have lent themselves very well to the shift to remote and hybrid.

We are also seeing in other parts of the economy, which people may not think can plug in to hybrid immediately, some really interesting shifts. Plumbers, car mechanics and other types of professions are using technology like Zoom to increase efficiencies and productivities in their work as well.

Q68            Chair: Let me just unpick that. If I need a plumber, presumably they will dial in on Zoom and give me a diagnosis before turning up and charging me a call-out fee. Is that what you mean?

Charlotte Holloway: Yes, we have seen some examples of small businesses, such as plumbers, doing that. Rather than doing the call-out and working out what bit you need—forgive me; plumbing is not my field of expertise—you can do an inspection of what the issue is. It saves time for the employer—as well as manpower—and for the customer. They can get a better idea of what the issue is. Using remote and hybrid technologies saves travel time and CO2 as well. They can communicate with their teams afterwards when they have that analysis.

Neil Ross: As someone whose boiler has just broken and who is trying to get it fixed, that is exactly what is being done.

Q69            Mark Pawsey: Neil, you told us that there seems to be greater adoption of technology generally in London and the south-east geographically. Could you tell us a bit more about the different sectors of the economy where adoption has been highest?

Neil Ross: Yes. The point about London and the south-east comes from the fact that the sectors of the economy that are more prevalent there are higher adopters of things like AI. For example, in sectors such as financial services, technology and communications there is about a 30% adoption rate for artificial intelligence technologies. For industries like retail and hospitality, that is a lot lower; it is about 11%. You find that is having the biggest productivity-boosting effect in those areas.

Q70            Mark Pawsey: Are there any things we can do to increase rates of adoption in some of those sectors? Hospitality is difficult because somebody has to cook your meal and somebody has to bring the plate to your table. How can technology help in those areas where there has been low uptake?

Neil Ross: Ultimately, investment in technology is an R&D expense. It is helping the companies in those areas invest more in research and development and those productivity-boosting technologies. You might look at grants, tax credits or how we use the R&D tax credit to support companies of that size. That might be one measure by which you can do that.

In retail and hospitality, we saw over the course of the pandemic how the use of apps for orders or table reservations helped companies manage much more effectively. That has productivity-boosting effects. It is those sorts of things we want to be encouraging.

Q71            Mark Pawsey: Brian, Amazon is a pretty big retailer. Can you give us some examples of how you have used new technologies in Amazon to deliver your services?

Brian Palmer: Thank you for the opportunity to speak today. AI and technology more broadly present a tremendous opportunity to improve and elevate the quality of jobs. You mentioned the retail and logistics sector, where we are most prominent. When we deploy technology like this, we start from the experience of the employee and work backwards. Much has been made, and rightly so, about productivity, but for us that experience and improving safety in our facilities is the top priority.

Giving an example, we deploy Amazon Robotics in our fulfilment centres. It has been about a decade since we first deployed Amazon Robotics in 2012. Over that time, we have seen tremendous job growth around the world, including here in the UK. We are deploying technology on the hardware siderobotics, automation or automated conveyance. On the software side, we have seen a reduction in the safety issues that usually present in the logistics industry. These include musculoskeletal disorder, which would be a muscle sprain or something like that.

By deploying this technology alongside our people, working with them directly to understand how it is working and taking their feedback to improve on the fly, we are able not to replace employee experience but to create a better and more elevated employee experience.

Q72            Mark Pawsey: Tell us a bit more about your use of algorithms in that process.

Brian Palmer: Thank you, Chair, for allowing us to venture outside of AI, which is a broad term to describe software. In the algorithmic space, we are usually deploying this technology essentially to monitor the health of the network and to engage in inventory control or quality assurance, like any retailer would.

When we deploy this technology, we focus on three principles. The first is—my colleague has mentioned this—transparency. It is vital that there is an open dialogue with employees on what technology is being deployed that could impact them and that we can take feedback as a two-way dialogue.

The second is the importance of technology with purpose rather than tech for techs sake, if you will. Again, we are working backwards and improving their experience, not just deploying software or hardware again for techs sake.

The final thing—this is very important—is the human leadership element. People are the foundation of Amazon and always will be. When we deploy things like algorithmic software, it is vital that a human is always in the loop and that human leader can review decisions. Some decisions should only be made by people. We do not see that human leadership element going anywhere anytime soon.

Q73            Mark Pawsey: Charlotte, you said the UK had a decade of adoption in two years and had adopted new working practices and procedures more quickly than other areas. What do you attribute that take-up to? Sometimes we are told that we are slower in some of these things and we are resistant to change. Your evidence seems to contradict that.

Charlotte Holloway: Thank you for the question. Much of this was down to the pandemic and the pressures it put on significant parts of the economy.

Q74            Mark Pawsey: The message I was getting from you was that other countries did not adapt as quickly as we did.

Charlotte Holloway: Yes. We saw that SMEs in particular had a higher adoption rate of technologies like Zoom than other countries, including the United States, France and Japan.

Mark Pawsey: Yes. I am interested to understand why.

Charlotte Holloway: In part, it is because some of the conditions in the UK mean we have quite a savvy consumer base. People are quite advanced in their use of smartphones compared to other countries. I am sure techUK will be able to speak to this point. That meant people were able to use some of those skills they had in their personal lives and take them into the workplace or into the new remote workplace. Just as people would be using technologies to, say, contact their family, they were then applying those in the workplace as well. It was a very quick shift. We saw different types of practices working in different environments and different companies.

There are some underlying conditions that mean we are good adopters of tech. We are good consumers of tech, as a country. For the first time, we were able to get those benefits from how some of the small business community and, indeed, the wider business community were adopting it.

Q75            Mark Pawsey: Andrew, if I could turn to you, as somebody who represents the workforce, that is quite a positive thing to hear. We often hear that workers are sometimes resistant to adopt new practices. What is your view on what Charlotte is telling us about how we are quick adopters of AI and new technologies?

Charlotte Holloway: If I can just clarify, Mr Pawsey, I mean the use of remote and hybrid technologies. The UK does still have certain types of digital skills challenges. I would not seek to undermine that. I am talking about the shift towards going online to be able to conduct the day-to-day interactions of work.

Q76            Mark Pawsey: Yes, but you are telling us that we have moved more quickly than other places. Many people would be surprised by that. I am interested to hear Andrews view on how the workforce has adapted to that change.

Andrew Pakes: It is a very good question. There was wholesale adoption in many industries and job types, particularly during the pandemic. As I said earlier, we were able to keep people safe, connected and broadly working only because of the deployment of technology at rapid speed. In that speed, there is a risk that we have missed some of the disbenefits and problems of it.

Q77            Mark Pawsey: Tell us what the disbenefits are.

Andrew Pakes: We often talk about how we now have the ability to take our work anywhere for any job. That also means our work can follow us everywhere. That balance between connectivity and human agency, the right to a private life, transparency and knowledge about what is happening to you can be very difficult.

One of the big problems is that companies have adopted this technology at high speed, but transparency, in our opinion, has not always followed through. A good example would be the adoption of platforms like Teams. In the UK, our economy is of such a size that most of these products are designed in other parts of the world. Most companies will not be designing their own internal management tools; they will be buying them off the shelf from other companies. At the point something is introduced, that is really useful. We all enjoy and we are all frustrated on occasion with the likes of Teams. Say a productivity scoring tool is introduced into that. If, as we have found, six months down the line people are holed up in disciplinaries and managers are saying, “We have been looking at your email traffic, the software you have been using and your websites; we do not believe you are a productive worker, that gets into the creepier use of this technology. But it is not the technology. This is why your inquiry is really welcome. It is the management practice and how technology is applied that we need to fix.

Q78            Mark Pawsey: You reference Teams. We do not have Teams with us, but we do have Zoom. Are there any challenges in the programmes Zoom has introduced that have caused you some anxiety?

Andrew Pakes: Not that I am aware of, no. Zoom is a connection tool that allows people to connect. There are a range of these omnibus tools now. There are products on the market globally that sell themselves as being able to track your workforce wherever they work.

Mark Pawsey: Zoom are good guys in this.

Andrew Pakes: Largely you are good guys, are you not?

Chair: Charlotte agrees.

Charlotte Holloway: Zoom does not market itself as a tool in the way that has been implied by the question.

Q79            Mark Pawsey: Andrew, you have told us about that. Are there any jobs that are under threat as a consequence of the adoption of some of these new technologies?

Andrew Pakes: There is considerable concern among the members we represent around work intensity and their sense of respect or agency over the work. Again, I would give you the example of task allocation software. When I turn up to do my task, I log into my tablet or computer and it will tell me what the task is and how long I have to do it.

On one level, that could be positive. It is a connected and easy way to get that. People are worried about where that data goes behind the scenes. It could be used to create a league table of workers so you can suddenly say, “Mr Pawsey, you are a slow worker. We think you should work more quickly. It is the use of it by management processes and people analytics.

Q80            Mark Pawsey: Is there evidence that is happening?

Andrew Pakes: We have seen that in a number of industries we represent.

Q81            Mark Pawsey: What type of industries have presented problems, in your view?

Andrew Pakes: Our members are seeing products coming into engineering roles. These are skilled professional trade type jobs. There is a widespread issue around this in what would be regarded as more traditional manual collection jobs. This software is now being applied in a wider set of skilled roles. That professional pride that people have is now being challenged by the lack of transparency.

Q82            Mark Pawsey: Brian Palmer, I read in the financial pages that Amazon is looking very closely at its headcount right now. Can you tell us a little bit about how that arises and the link to AI and technology there?

Brian Palmer: Like you, I have seen the press reports. That is not something I am in a position to comment on today, but I will tell you that in field operations, which is my area of expertise, we continue to see growth alongside technology. It is not a zero-sum game. As I mentioned, we introduced Amazon Robotics 10 years ago, with software alongside it to guide that process. During that time, we have introduced over 1 million jobs, including 70,000 here in the UK.

Q83            Mark Pawsey: How much more efficient are you today than you were 10 years ago? For example, how many more orders can be picked and how many more deliveries can be made as a consequence of the adoption of the technology you are telling us about?

Brian Palmer: I do not have a specific productivity figure for the UK that I can reference. The job growth in our Amazon Robotics facilities is still significant even alongside what we call our traditional facilities. It is not a zero-sum game where robotics facilities employ fewer people.

We are deploying the tech. Yes, it does improve productivity, but it also makes our people safer and improves their quality of life. As my colleague to the right mentioned, we are continuing a transparent relationship with people in terms of what is expected from them and what they expect from us as providers of tech. The feedback we get from our Amazon employees when we deploy technology has been very positive.

Q84            Mark Pawsey: Neil, you referenced the issues with participation in the labour market and the productivity of UK workers. Are there any jobs that are at risk or are we simply managing to do as well as we can with the limited number of workers we have? Is that how AI is helping us most?

Neil Ross: Technology always ends up replacing roles. There are lots of communication roles that have been replaced over time by different forms of technology or the internet etc. The key thing is to ensure that, as new technologies develop, we train people to use them and create new roles off the back of them.

There is really good international evidence to suggest that higher levels of technology adoption within companies prevent job losses. If you imagine a job as a collection of tasks, the more you are able to encourage high tech adoption across those tasks, the less likely they are to be replaced and the more likely you are to find new ones. If we can continue to invest in the workforce, training and technology adoption, we should be able to avoid a loss in the labour market.

Q85            Charlotte Nichols: Mr Pakes, from what we have heard so far, it is very clear this is not binary: “AI good or AI bad”. There are lots of ways in which it works within the labour market. Within the companies in my constituency that Prospect represents, such as Sellafield, we have the robotic snakes that go into the reactors at nuclear power stations and things like that. Certainly, those are things that you would not want a human worker to do.

The part I am interested in from your point of view is where there needs to be action to ensure that automation is not reinforcing or worsening inequalities within the workplace, when it is potentially being used for bad.

Andrew Pakes: You are absolutely right. There are some really great examples from the energy sector in your constituency, where we represent members. We also represent members in the likes of the Met Office, who are using AI technology and supercomputers to help make our weather predictions better. We represent members in air traffic control, who are using more powerful supercomputers to help improve aviation safety.

There is also the other side of it, and this is where we think the focus will be. What sits behind AI and this technology is the use of our data to make choices about individuals working lives and where they sit within the workplace. Many of our laws are currently based on a physical presence. Health and safety is about physical harms and risks. We do not yet have a language or a legal framework that adequately represents harm or risk created by the use of our data.

If you follow the data lifecycle, you now have AI and algorithms being used in recruitment software. It is used in the production process and how people are managed. It can be used to hire and fire us. There is an example we have been dealing with recently, which is not AI but digital. We saw how P&O was treated, but we have seen a very similar digital P&O in how our members at Twitter have been treated in recent weeks. People were communicated with by email without any human contact at what you would hopefully agree was the point of maximum anxiety, when people know redundancies are being created.

“Human in the loopsounds like a really good term, but it can cover minimal action as well as good management practice. We would argue there should be a set of data rights from day one in the workplace to ensure people have that transparency so we can define some of the concepts around what transparency looks like.

We would like to see accountability as well. In this day and age, it is not the right answer to be able to say, “The algorithm did it. I would probably say that most managers and most of our members would not understand the algorithm itself, but someone needs to be held accountable for it. In the same way, if someone has an accident in their workspace, there is an accountability process in that workspace. Where is the accountability if an algorithm makes the wrong choice about someones life? How do we ensure people get some accountability within that?

The same would come with surveillance. The survey work we have done and wider polling show that one in five British workers are now subject to some form of surveillance software. Largely, the legal framework for that comes through GDPR. We would argue that individual privacy rights are not the way to deal with a workplace issue. How do we update our employment framework? It is about rights and balances. We think techUK and other colleagues have really good conversations about this. We have laws that are not fit for the way in which digital technology is transforming the work relationship and how people are managed.

Q86            Charlotte Nichols: When it comes to saying,The algorithm did it”, there is not yet within our legislative framework an understanding that the algorithm is only as good as the information that is put into it. In the case of Twitter, they laid off quite a lot of their programming staff because they asked, “How many millions of lines of code have they generated?” and got rid of the bottom 20%. The people at the bottom are probably dealing with more complex cases, for example. That question did not say whether they were good or bad.

On that, Mr Ross, what work needs to take place by the Government to support the kinds of things Mr Pakes was talking about, so that we have rights and balances? Companies need to be able to adopt AI, but we need to make sure that it is being used responsibility, that there are rights of redress and that requirements around transparency are met.

Neil Ross: We work very closely with Prospect and others on an annual digital ethics summit that we run, trying to examine these issues. Andrew is right in saying that the legislative framework is out of date. The Government have missed a number of opportunities to introduce an employment Bill to update employment rights alongside the changes in the workplace we have seen.

Whether you are in the gig economy or other parts of the labour market, that means you end up having to rely on quite outdated structures of how the labour market works, which are not fit for the post-pandemic world we live in today.

There are therefore two ways in which to approach this. As Andrew said, in the GDPR there is something called a subject access request, which is useful for gathering data about what information a company holds on you, how it is being used etc. That is a right that is available to every UK citizen. We are glad about that, and we are keen to see that continued in the Governments reform of the GDPR. That is something we advocated for.

On the other side of things, when it comes to AI, the Government are currently developing a policy framework for artificial intelligence technologies. We saw a policy statement come out over the summer and we are expecting a White Paper by the end of the year. There are six broad principles they want to see applied to AI technologies across sectors.

If I can just check my notes, that includes ensuring that there is a legally defined person who is responsible for AI governance within a company where it is being used. You have to clarify routes to contestability and redress. You need to be able to explain to users and customers how the system operates.

There are other rights around making sure it is technically secure, which is about mitigating the bias and the cybersecurity issues you have raised as well as making sure the software is safe overall. We have been working with the Government very closely on their AI policy. We think they are basically along the right lines with those six cross-cutting principles.

Q87            Tonia Antoniazzi: I want to focus on workforce skills. I know we have touched on this so far. Neil, what upskilling and retraining is needed to equip staff with the necessary skills to work alongside AI, robotics and algorithms? Is the current investment and focus on digital skills sufficient?

Neil Ross: To take the second part of the question, is the focus sufficient? No. We did a survey of our members recently asking them a number of questions about how they are feeling in the economy at the moment. Broadly, they said that confidence had fallen. One of the main reasons was a huge increase in staff costs. That mainly points to the extreme competition for people with those digitally skilled roles. As I was setting out at the beginning, we have a real supply problem here.

It is quite hard to predict the skills that people need because this is an emerging and evolving sector. It is a blend of creative, coding and technical skills that we need across the piece. Ultimately, we need more post-school provision. We need to reform the apprenticeship levy, ensure we have more short modular courses for people to take throughout their working lives and support people to see retraining as something you just do as part of work rather than staying in the same role throughout. It requires a complete rethink of our education system particularly post-16 and post-18.

The Government ran a task force on digital skills between DCMS and Downing Street under the last last Prime Minister. I am not sure how it is being picked up by the new new one. That is something we need the Government to refocus on as soon as they can.

Q88            Tonia Antoniazzi: It is quite interesting. When I was teaching, many moons ago, we used to say to the children that the jobs of the future do not exist now. I have heard it so many times. Reskilling and unpredictability are definitely there throughout your working career. It is about preparing people for it. You have talked about post-16. Are changes in the curriculum required now to ensure children are prepared for the future?

Neil Ross: It is not my area of expertise, but people would agree that there are changes you could make to the curriculum. We are coming up, particularly in England, to 10 years since the last major curriculum revision. There is potentially an opportunity to look at how we integrate more technology use as part of that.

Again, we do not want to rely too heavily on the technical side of things. It has to be a more holistic approach, getting people used to using different kinds of technology. The point you make is right. Andrew mentioned earlier that 50% of our 2030 workforce is already at work now. We really need to boost in-work re-training, as well as post-16 and post-18 retraining, to make sure we are ready for the jobs that are going to come around.

Q89            Tonia Antoniazzi: Charlotte and Brian, I would be interested to know what your thoughts are, at Zoom and Amazon, about these skills that are needed.

Charlotte Holloway: What is really interesting is that, anecdotally, we are hearing from a number of our customers—large enterprises, large established UK firms—that they are looking for the workforce of the future, those new graduates who are coming through, to be hybrid ready for the world of work.

A lot of our customers are telling us that they want to be able to get the best people for the role, no matter where they are across the UK. There is quite a positive story there about having certain roles not just in London and the south-east but spread across the UK, which I am sure we can come on to. They are looking for a workforce that is hybrid ready. That is about not just new people coming into the labour market but about the managers who look after them, who are comfortable managing people for part of the week or the majority of the week at a distance.

To Andrews earlier point, which was really well made, that is not just about the technologies; it is about the practices, the policies, the culture and the leadership. Tech can only be used for good in the environment it operates in. We are seeing employers want that workforce to be ready. That has to be a combined effort between the framework in which they operate and the leadership within those companies to get the skills balance right.

Brian Palmer: I would take this in two parts. There is the education and skills component. Earlier education and improved literacy at a younger age in both software and hardware programming and things like mechatronics or robotics are incredibly important. That is where we have focused our energy with the Amazon Future Engineer programme, in partnership with Teach First.

In an ongoing way, there need to be tools and resources available to the workforce to improve themselves where they see fit. For example, Amazons Machine Learning University and AWS re/Start, which is a certification and professional advising programme, are both available free to the general public. These are ways that we are providing new opportunity to folks who did not have the opportunity to learn in an earlier education forum.

The other point I would like to make here is about the ability of technology to improve accessibility for roles that traditionally would not be as accessible to folks, for example people living with disabilities. This is an issue that is very close to my heart. In our fulfilment centres, we have been able to use advanced technology, an AI-backed hardware like computer vision, for example, to allow folks who have low vision, blindness or partial blindness to perform tasks or roles they traditionally would not have been able to do.

The deployment of robotics expands the aperture of people who are able to work in the warehousing and logistics business, which of course traditionally has a physical element to it. When we take the heavy lifting out of that environment and make it more broadly available, we see that as a positive for engaging people who are just coming into the workforce or who are looking for a career change but also for people who can pivot that into a new career through Amazon Career Choice or outside of our operations.

Q90            Tonia Antoniazzi: Andrew, how are your members reskilling people already in work? You have spoken about the need to be dynamic, having transition plans and people having a future. That is really important. For us, how do we achieve the balance of ensuring the economy, the individual and the Government are all playing their part in reskilling people?

Andrew Pakes: The partners you mentioned should all be at the table coming up with the plan for how we get this right. If we are honest, productivity has been a long-term issue for the UK economy. Just inserting technology alone will not change that. It is about the management culture, the business model and the wider economy, which is why it is absolutely right that this Committee looks at this.

Too often, digital and AI is looked at as purely a DCMS issue rather than an economic issue and a labour market issue. We need industrial strategy. For me, if there is a message to take out, it is that, like we have a just transition in energy, we need a just data and digital transition into the new economy and skills. If we do not get this right, the risk is that we are not going to have high innovation and high involvement.

When it comes to the future of the UK economy, what our members want, and what GMB and other unions want, is high involvement for the workforce, with people on the shop floor or the digital shop floor who understand that. The trouble the UK has—this is one of the weak points inside that lack of industrial strategy at the moment—is that adult skills and retraining are largely an individualised pursuit. If you compare and contrast that to Germany, Sweden or other areas, they have a higher level of social partnership, recognising that we have an economy-wide responsibility to enable individuals to retrain.

This should not just be about an individual taking up an online option or doing something in their spare time. This matters to all of us. Employers are largely incentivised to train people for the job they do and maybe the next iteration of it. Who has the responsibility to train people for jobs that do not yet exist? That is the balance for people in work. Let us take a member I represent, who may be approaching my age of nearly 50 and who is mid-career in engineering. What incentive or economic ability do they have to step out of the labour market and enter into some new form of training to find a new job at the end of that?

Two weeks ago I was in Sweden looking at the work employers and unions are doing together on their job security council. That is primarily aimed at getting those people mid-career to have both the support and careers advice and the economic ability, supported by employers and unions, not just the Government, to do a career switch. Their whole economic model is based on retraining people in the job market at the moment to do the future jobs. We do not have a current plan in the UK to enable us to compete with that.

Neil Ross: I would completely agree with Andrews points on industrial strategy. For tech sector jobs, there is effectively a market failure in terms of matching supply to demand. There is a huge amount of work on this being done by companies like Zoom and Amazon, and across techUK membership. They are spending vast sums of money on retraining programs, but there is no focused conversation with the Government to channel that resource and effort into the right places. That is something we really need to engage on.

I am quite heartened by the new Education Secretary. She gets this issue, but we need to see a lot of movement in Government on it. We have not been doing enough quickly enough.

Charlotte Holloway: If I could reinforce that, I am a techUK board member. I know that techUK has been making the point Neil just raised not just over the last year but over the last decade and more. That has been a consistent message to Government from the sector. Yes, I can see that failure.

Q91            Andy McDonald: If I could direct some questions to Brian, I would be grateful. Have you carried out any analysis as to how automation has benefited productivity? Can you also perhaps tell me how you make sure the benefits of that productivity are shared with your workforce?

Brian Palmer: Again, the use of automation and robotics in our fulfilment centres is not designed to replace or automate out jobs. We target mundane, repetitive, physically or mentally difficult tasks.

Q92            Andy McDonald: Has it done that? Are jobs now done by machines and programmes that were once done by human beings? That has presumably happened, albeit not the design.

Brian Palmer: Yes, I would suggest tasks within a function might be done. There are over 60 human-led functions in a given fulfilment centre. There has been some press about the number of packages that will be touched by automation getting closer to 100%. What I would suggest to you is that the number of packages that are not touched by humans will continue to be at 0%. This is still a critical part of the value chain.

In terms of improving outcomes for people, we see an improvement in safety, the reduction of things like repetitive motion injuries or musculoskeletal disorders and improvements in employee retention. The jobs are more sustainable. Candidly, logistics, warehousing and transportation has been subject to high attrition, but we continue to perform better than the industry on that.

My colleagues mentioned the importance of transparency and the mental or physical stress that can be present working alongside technology. When we have a candid conversation and an open dialogue with our people, and we take feedback from them and change systems based on that input to improve how the systems work with them, it serves to mitigate a significant part of that.

Q93            Andy McDonald: You employ 70,000 people, you said.

Brian Palmer: That is right.

Andy McDonald: Are those people directly in your employ or are they people who are gig workers or platform workers? What is the proportion, or is it on top of that?

Brian Palmer: All 70,000 people I mentioned are directly employed. The majority of those will be in our fulfilment operations.

Q94            Andy McDonald: They are directly employed, but your business model depends upon an army of people delivering the goods for you who are not directly in your employ. Is that right?

Brian Palmer: That is correct. You are referring to the last mile delivery component. We work with a range of transportation and delivery companies, including big companies like Royal Mail or multinational ones like UPS. We also work with our delivery service provider programme, which is small and medium-sized delivery and courier businesses that leverage Amazon technology, but those drivers are not employed by us.

Q95            Andy McDonald: The last mile, the last bit upon which your whole business model depends, is outwith your responsibilities albeit totally within your control. You control every element of that persons job. I have been out with an Amazon driver. I was very interested in your remarks around transparency, tech with purpose and people being at the centre of it. Are those people not at the heart of the group you are thinking about? Are they not your concern?

Brian Palmer: Of course they are our concern. Our delivery drivers are working through delivery service partners, who can choose to use Amazon technology or not. We provide tools like routing software or dynamic systems that allow them to make changes on the fly. That provides a support line for those drivers directly back to us.

Q96            Andy McDonald: On that very point, the driver I was out with was very frustrated, not only by having to meander through the lanes and narrow alleyways seeing DPD and Hermes drivers coming the other way and clogging up the roads, but also by the programming that Amazon had delivered to him. He was not able to have any input to say, “This is wrong. Can you please listen to me? I do know what I am talking about”. There was no opportunity; certainly nobody was listening to him.

Given all of those responsibilities that you set out, I would be very interested to know where the trade unions sit in this. Where is their voice? You are talking about the productivities you are securing through these processes, but I would come back to my original question about the benefits being derived for working people. You have responsibilities under ESG, with which you will be very familiar. How are you discharging those responsibilities if you do not directly employ the people you ultimately depend upon?

Brian Palmer: First of all, let me start by saying that I am sorry that you had that experience. Generally that is not consistent with my time out on the road for Amazon, which I have done both here in Europe and in the United States. The mechanism to provide feedback to Amazon Logistics, which is the group you are referencing, is via either the delivery service provider directly—that is usually collated by them—or a 24/7 hotline that they also have available to them. They can always talk to a human. That ability to engage a human and provide that direct feedbackhaving a human in the loop to make those decisionsis vital.

We mentioned, “The algorithm did it. For us, that is never a good enough answer. These are decisions that have to be made with a human leader who can review and understand the nuances of an individual either in our fulfilment operations or out on the road. They have to have those individual experiences and the human empathy and leadership that is so important.

Q97            Andy McDonald: I am struggling to understand how you can apply these standards of behaviour, but you are not even in the room. If you are not in that relationship of employer-employee, you can abdicate your responsibilities for that, and your business depends upon it. Is there not a direct contradiction in that model?

Brian Palmer: I do not. We are very proud of the delivery service provider programme, which has allowed thousands of small delivery companies to grow their business, employ new drivers and grow their workforce.

Q98            Andy McDonald: Moving on, a big issue that concerns people in this space is the development of technologies around surveillance. We have heard a lot said about that thus far. There are a couple of points to this. I did not really think I got much of an answer to my questions about trade union recognition, whether there are collective bargaining arrangements going on and how widespread those are. I would be interested to hear from you on that.

Does Amazon deploy unionisation heat maps, as has happened in the United States, to see where there is pro-union sentiment? Do those technologies exist within Amazon in the UK? Are those being deployed here?

Brian Palmer: No. I would preface that by saying I am not an employee relations expert in the UK. The answer is no. We are not tracking union activity. The heat map you suggested is not something that is deployed here.

Q99            Andy McDonald: That is good to know. Just pushing a little bit on the surveillance issues, the DCMS Committee pursued this question very recently, earlier this month. Dr Matthew Cole from the Fairwork project, you may have heard, told the committee that, overwhelmingly, the evidence shows that the technologies Amazon uses are not empowering. They lead to overwork, extreme stress and anxiety, and there have been issues with joints and health problems. He went on to say that happier workers are more productive workers, but, when you have the level of surveillance and control that Amazon is capable of, it is less concerned with that.

We hear about peoples behaviours being tracked. I would like you to tell us a little bit about that. Ultimately, I want you to tell me who owns that data. This is about another human being and their behaviours, and decisions are made about their lives with the benefit of that data. Who owns it? Who has access to it? How are you responding to those challenges?

Brian Palmer: If I could take the first point regarding Dr Coles testimony, I will start by saying that we agree that unhappy, stressed or overworked workers are not a recipe for a productive workplace. We know—this is in our field operations and fulfilment centres, but this holds true throughout Amazon—the best person in an operation is somebody who is happy in their job, is engaged, sees a future there and is healthy.

What is even better than that person is that person backed up by great technology. That is why we deploy technology in fulfilment centres, for example, that reduces the need to manually scan items, where we can do that with something like computer vision. We are not seeking to monitor or surveil our people. Their privacy is something we respect. We do have things like CCTV at the entrance to the building, but when someone is working in a fulfilment centre, when they are at station, the focus of the software and hardware we have been discussing is on the goods; it is not on people themselves.

On your point about data—this is another very important point—we do collect and keep data that is legally required, such as next of kin and employment information and things like that. We also do keep things like performance data. We work with people and identify areas where we need to improve. That data is accessible to those people, and they can request that from us or see it through our internal systems. We have online tools that are made available to every single employee. They all have access to that data.

Andy McDonald: Thank you, Chair. That is probably enough from me.

Q100       Chair: Can I just check something? You just said to the Committee that the collection of data was based on your goods, not your people. That is not quite right, is it?

Brian Palmer: I am sorry.

Chair: You just said to the Committee that your collection of data, presumably in warehouses, was focused on the goods, the boxes of stuff you are shipping out, as opposed to your people, the people who are packing things and moving them around. I just wanted to check that. That is not actually correct, is it?

Brian Palmer: It is correct. The primary purpose of the systems that are deployed would be things like monitoring the productivity of the network and the health of the network to ensure we are planning correctly and matching capacity and demand so that even at the busiest time of the year the fulfilment centre will look as it does in July or March.

Secondarily, we use these systems for things like inventory control, like any retailer would. On your point about looking at performance, that is not where we focus when we deploy these systems. They are not primarily or even secondarily to identify under-performers. Performance-related feedback is really focused on safety and what we would consider internally quality, not on things like speed or productivity.

Q101       Chair: Let me give you an example. I represent a constituency that has one large Amazon fulfilment centre and another one just on the border. A couple of years ago now, I had a constituent who was 63. He was working with you to fill his time before his pension. He came to see me in my surgery. I do not know whether it was a wristband, a tablet or a device, but he had some kit that basically told him he was not packing things quickly enough.

He had had two strikes already. He had to go and see his productivity manager. He went to see the productivity manager, and the productivity manager said, “Why are you not packing things more quickly? The system is telling us you are not being productive enough”. He said, “I am 63. I am working as quickly as I can. It is pretty cold in here. I cannot go any more quickly”.

If you were to take that example outside the technology discussion, you would have some real concerns around age discrimination. This is your technology tracking that workers productivity in packing. That is surveilling the worker and not the goods, is it not?

Brian Palmer: I am not familiar with that particular example.

Chair: I do not want you to comment on the individual case because that is a cop-out of an answer. I want you to answer the question based on the example. Do you track the productivity of your workers in your warehousesyes or no?

Brian Palmer: Yes.

Q102       Chair: You have just said to the Committee that you do not track the worker. That is not true, is it?

Brian Palmer: What I said was that the systems are not focused on tracking individual workers—

Q103       Chair: What do you mean by focused”?

Brian Palmer: Their primary purpose is not to track individual productivity.

Q104       Chair: From the workers perspective, if they have a bit of kit or a device that is saying, “You need to be quicker” all of the time, that feels like the focus is on them and not the goods. Do you recognise that?

Brian Palmer: We do. You said that this was a couple of years ago. We have made changes in response to employee feedback on how those metrics are assessed. As I said, early in the tenure when you are starting out at Amazon productivity is not a factor in performance evaluation. We focus on safety.

Q105       Chair: He told me that he was going to be fired if he had three strikes on his productivity tracking. Are you saying that is not Amazon policy?

Brian Palmer: The folks who leave the company for performance-related reasons are a very small minority. That is not where we focus. If someone is having performance-related issues, we want to understand why and how we can help them. Maybe that is not the right function for them to be performing within the fulfilment centre.

Q106       Chair: It is really important that you answer the answer I am asking you. If someone has three productivity flags on the system, can they be fired at that point?

Brian Palmer: Yes.

Q107       Chair: What you just answered when I asked you that question before was something entirely different. Now you have given me the actual answer to my question, do you not recognise that is a problem? There is an algorithm tracking a human worker. The computer assistant is saying, “This person is not quick enough”. To be fair to him, the guy was 63. You were going to sack him based on whether the algorithm decided he was productive or not. That is an issue, is it not?

Brian Palmer: To be clear, the algorithm is not making those decisions.

Q108       Chair: No. The conversation he had with his productivity manager was, “Well, the system told me you are not being productive enough”. That is not a human in the loop; that is a human telling another human what the computer told him.

Brian Palmer: The important thing is that this is an ongoing conversation between the manager, onsite human resources, which are there, and the employee. Where we identify issues with performance, again, we would recommend performing a different function within the fulfilment centre. As I mentioned, there are over 60. I worked in fulfilment centres early in my time at Amazon.

Of course, we set performance benchmarks based on similar groups of people. They are not constantly ratcheted up. That is not a practice that we engage in. It is self-defeating and uneconomical to create an environment where people feel like they are constantly being ratcheted up. We have to continue to work with our employees and identify areas where we need to make changes and where we can improve.

Again, as I said, if the issue is performance at a specific function, the vast majority of the time that employee will be asked to perform a different function or they can go to HR at any given time and raise these concerns directly with them.

Chair: My constituent did not feel like that, I am afraid.

Q109       Charlotte Nichols: Mr Palmer, documents filed by Amazon with the US National Labor Relations Board revealed that between September 2017 and August 2018 about 300 workers were let go at its Baltimore warehouse for what the company called productivity reasons. This was more than 10% of the workforce.

The companys lawyer wrote that, “Amazon consistently terminates fulfilment centre associates for failing to repeatedly meet the standardised productivity rates”. Was that productivity rate based on AI or other forms of algorithmic working?

Brian Palmer: No.

Q110       Charlotte Nichols: Amazon extensively uses AI for other purposes. Why do you not use it for warehouse performance management? Is that not at odds with the answer you just gave to our Chair moments ago?

Brian Palmer: No. The software that is deployed in a fulfilment centre is not designed to set benchmarks independently or automatically. These are benchmarks set based on achievable performance levels, based on historical performance, that are reviewed by human leaders before they are put in place. It is not independently automated from a decision standpoint.

Q111       Charlotte Nichols: Analysis shows that November is the worst month for ambulance call-outs to Amazon fulfilment centres. There is almost a 50% increase between October and November every year over the last five years. It seems to me, with an Amazon fulfilment centre in my constituency, that there are human beings being pushed beyond the limits of endurance by productivity targets set at least in part by robots.

You worked in a senior role for Amazon in the Washington and Baltimore area while the layoffs I discussed earlier were taking place. Do you feel personally conflicted about more than a 10th of the workforce being laid off for productivity reasons?

Brian Palmer: I did not work in the Washington area at that time. I was based in Seattle in 2017. I am not familiar specifically with the case that you are raising.

What I can tell you is that performance-related decisions are a very small minority of people who leave the company. Again, it is not a decision that is made in real time. It is not as if you have had a bad day and that is what the computer spits out. This is a discussion that is ongoing with the human manager to identify areas where you can improve performance.

We recognise that warehousing jobs are not for everyone, but through the deployment of technology we have been able to accomplish an environment that is significantly safer and has significantly better retention rates than the industry.

Q112       Charlotte Nichols: It is clear from what I said about ambulance call-outs at UK Amazon fulfilment centres compared to other similar-sized logistics firms within the UK that the idea that your practices are creating safer workplaces is up for dispute.

In any case, based on the discussion we have been having here and the well-publicised labour problems within Amazon in the UK—including the spontaneous walk-outs earlier this year caused by concerns over safety—and given that we have GMB on the next panel representing workers in Amazon warehouses and career services, and Prospect on this panel representing tech workers, might it be time that Amazon recognised a trade union and worked with trade unions to resolve some of these issues?

Brian Palmer: Again, I am not an expert on employee relations here in the UK. We respect the right of our employees to join a union. We have always supported that right. At this time, no union represents an Amazon site in the UK, but we are entirely supportive of the right of our people to join a trade union.

Andy McDonald: Your respect for somebody’s right to join a trade union is immaterial. It is their inalienable right. Your respect of it is irrelevant. It is your ability to recognise trade unions and their ability to bargain with you on a collective basis. That is the key point you are not answering.

Brian Palmer: Sorry?

Q113       Charlotte Nichols: What my colleague is saying is that workers in the UK have a right to join a trade union immaterial of whether that is something that Amazon supports or does not. Our question is about whether Amazon should take the proactive step to recognise those trade unions and to collectively bargain with them in order to resolve some of the issues that we have been talking about today with Amazon and the way that technology is deployed within its workplaces, in the hope that some of these issues might be actually resolved.

Andrew Pakes: We would welcome that.

Brian Palmer: I am not an expert in UK employment law, but where we can work directly with our people, which is what we have been doing, we will continue to do that. If a workforce or a site is represented by any trade union, we are happy to have that conversation. That is just not the case today.

Q114       Charlotte Nichols: I am sure the GMB colleagues on the next panel will be very happy to hear that you are pleased to talk with them in future, as I know there are a number of outstanding issues at your warehouses in the UK.

Moving on from Mr Palmer, Ms Holloway, companies like Zoom have become much more well-known through the wider use of their technology during the pandemic. As has been mentioned previously, some of our hybrid working arrangements, even within this place, have relied on Zoom. How is your company continuing to evolve and promote new technology in the workplace as more people continue to work remotely or in a hybrid capacity?

Charlotte Holloway: Thank you for the question. It has been a great privilege of Zoom, the company I am representing, that we were able to work with the Parliamentary Digital Service during that time and with occasions such as the recent Zelensky address as well.

Zoom’s platform has been predominantly known by this place and by the broader population for its Zoom meetings function. As a platform there is a lot more, and we have new innovations coming live all the time. I could give you a long list of the various features that we have, such as our Zoom Phone product and our Zoom Contact Center, which is working with large enterprise customers. We have a number of other aspects as well as Zoom Events. A wide offering, as you would expect, for a growing challenger company to be looking at the new offerings it can make for hybrid work is the new normal for the vast majority of people in this country and across the world.

We are constantly looking at more things. We have Zoom translation tools; for example, you can translate conversations in real time on Zoom into Ukrainian or French. We are constantly looking at ways in which Zoom as a platform helps people to connect in the new world of work, but in wider life as well.

Q115       Charlotte Nichols: You mentioned in your written evidence that technology to support remote hybrid working has opened up the job market geographically, and has the potential to level up poorer or more remote parts of the UK. Again, certainly within my constituency, as a town between two major cities in the UK, we have noticed an increase in footfall in our town centres over the course of time where people have not been commuting into Manchester, Liverpool or London. Can you tell us a bit more about what these changes in the labour market might look like and what further changes you think we can expect in the future as some form of hybrid, flexible working potentially becomes the norm?

Charlotte Holloway: It is a really important question as we think about the world of work and the future of the labour market. It is broadly an untapped political and policy area for discussion as well, so I am really glad the Committee is thinking about it.

We released a report earlier this year that showed that, in the top 25 places that have seen a growth in remote working, or jobs advertised as having a remote element but geographically based, 10 were in so-called red wall seats across the country. The top one is in Worthing. Places like Burnley, Stoke and others are featuring very, very highly on that list. Indeed, the Chair’s city of Bristol features as a top 10 remote working hotspot as well. You are absolutely right to say that this has some really quite profound implications for how work can be based in different parts of the country.

It is also very interesting about how hybrid work is enabling more diversity or different types of people into the workplace. Women in the workplace have increased over the last couple of years, we believe in part due to those hybrid technologies. Women in work with children under three have increased by over 5% during that period. Two-thirds of women who have been surveyed recently—I think by Demos, but I will have to check—said that being able to work from home or remotely has enabled them to advance their careers.

For the geographical spread, we are seeing that right across different parts of the country. We are seeing really interesting examples in places like Hull, which has the “Work Hull Work Happy” partnership that Emma Hardy MP has been championing in this place. That is enabling businesses and locality to think, “If we can attract more jobs here with the remote element, how can that help regenerate the high street? How does that help have a wider knockon economic effect?”

Yes, there has been a big debate about what that means for the high street in large urban areas, but we have not necessarily seen productivity drops around it. There is a real opportunity. We have been advocating to the Government that they should adopt a hybrid working strategy. We are actually working with Prospect union and others on launching a new commission about the next wave of opportunities that can be seen across the UK. Because of our industrial makeup as a country, there are real opportunities here for all parts of the country to benefit from the growth this technology can bring.

Q116       Charlotte Nichols: I just have a question on the potential downside of this. I am very here for the utopian idea about what levelling up and hybrid working might look like, but we know that we see real geographic inequality with wages, particularly wages in London and the southeast being significantly higher than in other parts of the country. Do you think there is a risk that that regional inequality could become entrenched through hybrid working, when perhaps you are ostensibly based in an office in London but are living in a part of the UK where wages are lower? How can that risk be mitigated against so that we do not just trap people in other parts of the UK into lower wages, even though they are potentially working for employers in urban areas with traditionally higher wages? I would like to come to Andrew after this, for any thoughts he might have as well.

Charlotte Holloway: We know that there are impacts. There are types of jobs that would have predominantly been based in London and the southeast, but more and more employers are now open to having those roles in all parts of the country. You are absolutely right that that will have economic implications, which require further scrutiny. I am not here to eulogise that everything is wonderful. We are just saying that that is what we see and that is the result that is shown. The new commission that I mentioned wants to dig into some of those aspects.

We are seeing that about threequarters of jobs advertised as hybrid pay over the average salary, so there are gains to be made for local areas that want to think about having a hybrid offer, whether that is a coworking space or something like that, that enables those roles to be in other parts of the UK. Yes, when jobs and work happen in different places, it does have a knockon effect. We think there is a good regional distributive effect that can happen.

Back to Neil’s point at the start about filling skills gaps, you might have people who are prioritising family or certain things in their personal life, which means that they can have a role that historically might have had to exist in a certain urban centre, but now can be based in different parts of the country. We do not yet know the full picture but we know more needs to be looked at. There are positive elements as well as things that need to be looked at and mediated.

Andrew Pakes: We would see huge benefits. Many of our members got benefits of flexibility during the pandemic, but we understand there are negatives as well. It is about all job types being included in that. We represent a union where around half of our members attended workspaces daily during the pandemic because of the work they did in defence, in aviation, in national security and in broadcasting. Ensuring that more people get the option for hybrid working across all income brackets and job types is an element of inequality that does not often get talked about when it is about the death of the office. Not everyone works in an office.

How do we extend that level of flexibility? How do we make sure it is not just about working from home? It could be about termtime working hours. It could be about technology enabling a job share. It could be about other forms of flexibility that allow people to get the benefits that some job types have clearly enjoyed more of during that.

We would obviously say, “Involve workers in those decisions—nothing about us without us”. The easiest way to do that—and some of the great gains we and other sibling unions have seen during this last period—is where we have been able to collectively bargain, and where working time and working space have been part of pay terms and conditions. Many companies in this country are starting to do that successfully, precisely because they have involved their workforce in those decisions.

It comes back to the earlier point. We talked about “human in the loop”. The concept that we and other unions talk about is human connection, because human in the loop could just be the manager making the decision. That is a human in the loop. The human connection is individual workers and groups of workers having a say over that future to ensure we get the benefits on things like access and disability, but lose the dis-benefits of it.

The ONS figures before the pandemic were very clear. Home workers tended to be more women than men. They tended to work longer hours. They tended to be less likely to be paid a bonus and less likely to be paid as much as people in the office. The pandemic has smashed that. There are great opportunities through Zoom and others to radically change that, but we need to manage this carefully so that people benefit and places benefit, not just shareholders.

Chair: We have timed out, I am afraid. Thank you to all four of you for your contributions this morning.