The Government’s bold new plans - and even bolder use of language to ‘unleash AI’ - made the rounds in recent weeks’ headlines, but beneath the gleaming potential and the lofty optimism lies a critical question: will the rapid advance of AI lift UK workers as it claims, or leave them behind?
Drawing inspiration from Nobel laureates Daron Acemoglu and Simon Johnson’s Power and Progress, this piece by DG Cities’ AI specialist and Graduate Consultant, Nima Karshenas dives into the hidden risks of automation-driven displacement. By examining historical lessons and the blind spots in the unveiled AI policy, he uncovers how thoughtful procurement strategies can ensure that progress works for the many — not against them.
With AI investment growing at unprecedented speeds, governments are scrambling to stake their claim in this transformative market. The UK, under Labour’s growth-centric strategy, has every reason to push ahead. Early movers in AI have the potential to establish themselves in global markets and reap the economic rewards that come with it.
This pressure to act quickly has driven a liberal, growth-first approach. The plan’s emphasis on attracting investment, building infrastructure, and establishing initiatives like the AI Safety Institute reflects a strong focus on cultivating an ecosystem that supports the industry. But economic safety—the security of workers in the face of automation—remains largely absent.
Unpacking the Government’s assumptions
In his speech, The Prime Minister really drove home that the opportunity plan was going to deliver for workers, and this, through inspection of the corresponding policy document, is resting on the following assumptions:
AI drives the economic growth on which the prosperity of our people and the performance of our public services depend;
AI directly benefits working people by improving health care and education and how citizens interact with their government[1];
The increasing of prevalence of AI in people’s working lives opens up new opportunities, rather than just threatens traditional patterns of work.
At first glance, these ideas seem promising, but history tells a more cautionary tale. A discussion of the first assumption is where we first interact with and explore the concept of the ‘productivity bandwagon’ outlined by Acemoglu and Johnson in Power and Progress. The productivity bandwagon outlines a commonly accepted principle in economics that when there is a breakthrough or improvement in technology, this leads to increased productivity, that in turn translates into an improvement in worker conditions through wealth creation.
By examining two historical examples outlined in their research we able to take a more critical lens on this assumption:
The productivity bandwagon process illustration
The power loom era
The automation of weaving by the power loom displaced skilled hand weavers. While productivity soared, the resulting wealth accumulated among capital owners, not workers. Displacement without task creation left workers with lower wages, harsher working conditions, and limited agency for the next 60-70 years.
Illustration of the power loom at work. Source: Hulton Archive/Stringer/Getty Images
The digital revolution
Figure 3: Real Log Wages by education level in the United States (source: Autor, David. 2019. "Work of the Past, Work of the Future." AEA Papers and Proceedings, 109: 1–32.)
The rise of computers and automation starting in the 1970s promised greater efficiency, but for many workers (in the US) — and especially those without university degrees — real wages stagnated or even declined. The benefits of increased productivity were concentrated among the highly educated and capital owners, worsening income inequality.
This rise in automation was coupled with unprecedented neo-liberal tax reforms that were rooted in ‘trickle-down’ economics, that no doubt amplified income inequalities, therefore it’s difficult to directly attribute the fall in real wages to automation.
These examples reveal the key flaw in the productivity assumption: while technological advances drive productivity, they don’t guarantee better outcomes for workers. For that to happen, we must actively shape the conditions under which productivity gains are shared.
Creation or displacement?
The real question isn’t whether AI can increase productivity — it undoubtedly will — but what type of productivity we are fostering. Productivity that creates new tasks and industries can generate opportunities for workers. In contrast, productivity that automates existing tasks often leads to job displacement, pushing wealth upwards rather than spreading it across the economy.
This differentiation is critical.
The plan’s third assumption - AI’s ability to create new opportunities - recognises this challenge, but doesn’t address it head-on. The AI Opportunity Action Plan relies on market forces to create these new opportunities, ignoring the lessons of the digital age. Without targeted policies, there’s no guarantee the market will fill the gaps left by displaced jobs, especially under the deregulatory stance outlined in the plan.
The positioning the government has taken comes as a greater surprise, given the threats identified on lower-skilled jobs in the 2021 report by the Department for Business, Energy and Industrial Strategy - an analysis conducted on the back of Frey and Osbourne’s gloomy prediction in 2019 that around 35% of UK jobs were at high risk of being automated by computers. Although the estimated scale of the impact of automation is yet to materialise, there is no doubt that the recent rapid advancements in AI are going to accelerate this transition in labour demand, and the government needs an AI strategy that prioritises the economic consequences we can no longer ignore.
Nonetheless, by being conscious of the AI products we procure and develop, as organisations we can capitalise, excuse the pun, on the productivity that AI offers without displacing workers. The key message to drive home here is that as organisations, we need to procure AI products that Augment and Create instead of Trimming - but what does this actually mean, and how can this be built into procurement processes?
AI to Augment & Create (A&C)
What does it mean and what benefits does it bring to an organisation?
Now, whether an AI tool is augmenting and creating is entirely dependent on the context of each organisation, AI tools that automate certain tasks can in fact be augmenting and creating, seemingly a contradictory statement based on all that’s been discussed but leads to perhaps the most important distinction.
Every organisation must first take a critical look at their current operations and evaluate the impact automation will have on them. An example here to best demonstrate - if there are critical datasets that are held by your organisation, but could not previously extract the value from because of the extensive cost and resources attached to cleaning, sorting and structuring them, and AI tools can help automate that process at fractions of the cost, then you are bringing value to your organisation without trimming your operations. The distinction lies in understanding how AI interacts with organisational operations, as the same AI tool can streamline one organisation's operations while augmenting another's. It's not a one-size-fits-all solution, but rather a nuanced approach that requires a critical understanding of AI's role within the organisation.
A few examples of the kinds of tools we are talking about:
1. AI-Powered Data Querying and Insight Generation
AI tools can process complex queries across vast datasets, identifying actionable insights that support better decision-making. For example, local authorities might use such tools to analyse housing or transportation data, uncovering trends that inform smarter policy decisions. Similarly, businesses can employ AI to assess operational data, optimising strategies based on clear, data-driven insights.
2. Patient Health Summaries for Healthcare Professionals
AI can consolidate and summarise patient health records, providing doctors with concise yet comprehensive overviews of a patient’s medical history. This enables faster, more informed decision-making, improving treatment outcomes. Additionally, AI transcription tools can handle administrative tasks, such as updating patient records, freeing doctors to focus on seeing more patients and handling critical cases.
3. AI-Driven Sentiment Analysis for Public Engagement
Previously, robustly analysing public sentiment toward local plans or policies was challenging with standard techniques. AI now enables the processing of large volumes of feedback—be it survey responses, social media comments, or public consultations—to evaluate sentiment at scale. This ensures that community perspectives are integrated into the design and planning of local spaces, allowing for longer, more thoughtful, and inclusive planning processes.
Shaping AI procurement around augmentation and creation is not just a safeguard against workforce displacement, it’s a strategy for making organisations smarter, not just more efficient. This approach fosters a healthier work environment, supports long-term growth, and ensures institutional memory is preserved.
A smarter organisation: AI tools that augment decision-making provide workers with enhanced analytical capabilities, leading to more informed strategies and better long-term outcomes, improving productivity without compromising
A healthier work environment: Reducing repetitive tasks allows employees to focus on creative and high-value work, improving job satisfaction, fostering professional growth, and attracting high-level talent.
Long-term growth: Prioritising augmentation ensures businesses don’t just chase immediate efficiency gains but develop resilient, adaptable teams equipped for the future.
Institutional memory preservation: AI tools that work alongside employees rather than replacing them help retain and structure knowledge within an organisation, mitigating the risks of staff turnover, and an over-reliance on black-box technologies.
A&C Procurement Framework
Impactful procurement thrives on continuous learning and iteration, which we've embedded into a dynamic framework that integrates A&C principles.
A final note…
While we welcome the Government’s initiative in recognising the transformative potential of AI for the UK economy, considerable care needs to be taken in policy development to avoid repeating the mistakes of history. Economic growth alone does not necessarily lead to better outcomes for UK workers, and without thoughtful intervention, the benefits of AI risk further widening income inequalities and lowering real wages among UK workers.
A procurement strategy that prioritises AI tools which augment human capabilities and create new opportunities will not only safeguard the economy against growing inequalities but also deliver long-term, robust benefits for orgaget in touchnisations. While automation is not an inherent hindrance to the economy, understanding where and when to apply it is critical to the sustainability of both businesses and the wider economy.
At DG Cities, we help organisations navigate this evolving landscape, identifying, demystifying and implementing AI solutions that deliver impact on the ground, drive sustainable growth, whilst protecting the workforce. By embedding A&C principles into procurement, we can shape an AI-driven future that works for everyone. If you would like to continue this conversation, or get in touch about how we can help with your AI procurement, then please feel free to get in touch!
[1] We largely agree with this assumption, provided careful design of the AI products, but discussion is outside the scope of this blog.
