Manifesto Review: AI and Work

Craig Gent

25th June 2024


Introduction


With AI hype at fever pitch, the political mood music around artificial intelligence – and especially its consequences for work and society – has never been so consequential. The 2024 UK General Election, taking place on Thursday 4th July, will be the first since the latest wave of tools – from Chat GPT to AI art and beyond – have entered public consciousness. This blog surveys party manifestos across Britain and Northern Ireland, to assess the different legislative and regulatory paths any future governments (and oppositions) are likely to tread towards AI.

First, however, I offer some context to the major party’s latest policy offerings, by way of some significant recent treatments of AI by Westminster policymakers.

The political context


The unexpectedly early announcement of the general election forced a range of select committees to wind down completely during the ‘wash up’ period prior to parliament’s dissolution. This was the fate of the Science, Innovation and Technology Committee’s ‘governance of artificial intelligence’ inquiry, which hurriedly published its third report prior to disbanding. This, however, nevertheless provides a helpful snapshot of how UK policymakers were approaching AI just before manifestos were published, and Great Britain and Northern Ireland now head to the polls.

Perhaps unsurprisingly, the report cites the UK’s productivity crisis as a key reason why ‘public trust in AI’ is now a political priority, as well as the need to address some of the most familiar concerns – such as protecting intellectual property and tackling deepfake pornography.

Yet the inquiry’s most notable finding is its acknowledgment that AI ‘black boxes’ cannot simply be opened, and are not meaningfully knowable or even technically explainable. The report calls this realisation ‘paradigm-shifting’ as it upends a ‘reliance on explainability and understanding’ that has been well-established among legislators, policy thinkers and trade unions until now.

The report insists, therefore, that the priority of the next government should shift away from measures such as explainability or transparency to a focus on AI’s outputs, and ensuring they are fair, verifiable and not misrepresentative.

Yet if the AI conversation is to move away from scrutinising inputs – beyond the low-hanging-fruit of intellectual property – to assessing outputs, it raises the question of what exactly constitutes proper safeguards for a potentially epoch-defining technology. This is the context for gauging the main parties’ political instincts around AI according to their recently published manifestos.

The manifestos


The Labour Party

The 2024 Labour manifesto hopes that AI represents a growing sector in which the UK could emerge as a global leader. It promises to ensure AI is regulated and the public safeguarded appropriately. There is a recognition of the double-edged sword AI represents for workers, but a wider commitment to ‘digital transition’ backed up with promoting best practice around the application of technology.

The party aims for a quick win in banning deepfake porn, though how this would be enforced is left open. Labour aims to solve a real regulatory black hole with a wide-reaching ‘Regulatory Innovation Office’ in recognition of the cross-sector nature, and differing pace, of AI’s adoption. This represents a more general desire to regulate AI, and to introduce rights and protections for its users. Labour’s manifesto promises to implement in full its ‘new deal for working people’, which contains an explicit mention of safeguarding around surveillance, spyware and discriminatory algorithmic practices at work. The party also promises the ‘meaningful right’ for gig economy workers – who are generally managed according to algorithms via an app – to organise in trade unions. In terms of government procurement, Labour aims to secure AI-enabled CT/MRI scanners for the NHS – undoubtedly a good thing.

Greater reservations might be due for Labour’s desire for surveillance technology at work to be subject to consultation and negotiation. Although the policy may sound pragmatic, it risks ceding much of the conversation around AI surveillance to employers before negotiators, presumably trade unions, reach the table. Critics may note that such an approach sets up a ‘partnership’ dynamic that nonetheless accepts the premise of the employer’s right to surveilling workers – something that ought to give unions pause for thought. Labour also promises to remove planning barriers to new data centres, albeit with no acknowledgment of the immense ecological costs with which these are associated. Commentators such as the economist, and host of the Macrodose podcast, James Meadway have already warned of the plan being sold as local ‘investment’ despite creating few – if any – jobs.

A number of Labour’s other promises are left vague. The manifesto promises ‘data driven public services’ backed up by a National Data Library – but with little detail about how this would work and what it means for unprotected government departments. Labour says it intends to ‘protect good jobs and ensure good future jobs and that rights and protections keep pace with technological change’ – this is, of course, the sort of thing trade union leaders love to hear, but amounts to little by itself. Perhaps most consequentially, it remains unclear how Labour intends for the ‘regulation’ of the most powerful AI models to actually work.

The Conservative and Unionist Party

The 2024 Conservative and Unionist Party manifesto views AI as this century’s ‘steam power’ and sees the country’s future as being to spearhead the AI transformation in a similar mode to the industrial revolution. It intends to go big on processing power, and to apply AI to both the civil service (to cut waste) and the police (to boost law and order).

The Tories pledge greater R&D support for AI. Although likely to be rolled out within a typical business partnership framework, further research funding often has potential to reach a range of beneficiaries. The manifesto’s most welcome promise is protection and remuneration for creators – likely welcomed by creatives anxious about the future of their professions – though they may be confused about the caveat that the Conservatives want to do this “whilst also making the most of [AI’s] applications for creativity”, which could mean anything. Like Labour, the Tories want to procure AI-enabled CT/MRI scanners for the NHS – a good thing.

The pledge to equip the police with AI alongside new powers – the example given is facial recognition – appears despite algorithmic bias being one of the most widely-cited shortcomings of both image-based AI and the application of algorithmic decision-making to justice.

The Conservatives also hope to usd AI to further cut public services, though it is unclear what this looks like; a related offer to use AI to ‘free up’ doctors’ and nurses’ time also goes unexpanded.

The Liberal Democrats

The Liberal Democrat manifesto intends to make the UK a world leader in ethical, inclusive AI. Their manifesto says this is dependent on greater clarity, certainty, data privacy and accountability over the technology, and promises to look outwards to the global regulatory environment.

Like Labour, the Lib Dems want to work towards a cross-sectoral framework for AI. But by far the party’s most innovative idea is the use of citizens’ assemblies to guide, among other things, the use of AI and algorithms by the state.

Ed Davey’s party is the only one to promise a digital services tax increase for tech giants, but like many others, it views AI as the key to solving Britain’s productivity crisis. This is concerning because one risk of the AI ‘revolution’ (or bubble, depending on your opinion) is that workers’ rights and labour end up squeezed at the ‘growth and productivity’ pyre in the name of the greater good.

Seemingly at odds with the ‘bottom-up’ flavour of citizens’ assemblies, the Lib Dems want to negotiate the UK’s participation in the Trade and Technology Council with the US and EU, which is almost certain to aim to ‘regulate’ AI in close concert with transatlantic geopolitical and economic ambitions rather than the concerns of citizens.

As with the Conservative manifesto, the Lib Dems leave to the imagination the precise nature of AI’s role in saving NHS staff time and costs. And while the party promises to establish transparency and accountability for AI systems in the public sector, the manifesto doesn’t say how – or mention AI use in the private sector.

The Scottish National Party (SNP)

The Scottish National Party manifesto promises to make AI a priority for research and investment “to capitalise on the technological revolution and ensure its full potential is maximised”. Particular benefits are anticipated for public services, which the party – rather vaguely – wants to achieve in a way that is “open, ethical and transparent”.

Related to the future of AI at work, the SNP wishes to create a single ‘worker’ status for all but the genuinely self-employed – a move that is understood to address the current lack of protection for platform workers in particular, albeit unlike the Greens, the SNP does not mention platform workers explicitly. This measure is something Labour intends to ‘consult’ on, according to its ‘new deal for working people’.

The Green Party of England and Wales

The Green Party of England and Wales manifesto appears to be aiming to shore up its credibility on workers’ rights on the theme of AI and work with a raft of strong pledges. Its manifesto advocates a people-first ‘precautionary’ approach to AI, particularly in terms of its potential impact on workers, the climate, discrimination, liberty and creativity.

The Greens say they will foreground workers’ rights and interests where technology precipitates changes to working conditions, and talks up the need to protect artists and creators. The party says it will advocate full and equal employment rights for platform workers, with a right for employees to access their data and appeal management decisions – which are often made by algorithms. The Greens also want to address the issues of bias, discrimination, equality, liberty and privacy that have arisen from the spread of AI.

The less clear aspect of Carla Denyer and Adrian Ramsay’s policy direction is in its approval for aligning the domestic approach to AI with ‘neighbours in Europe’, UNESCO and ‘global efforts’ to support a ‘coordinated response’ – a sentiment also echoed by the Scottish Greens. While multilateralism is not necessarily a bad thing, the present state of EU politics should demonstrate that more cooks does not necessarily indicate greater protection of progressive values.

Plaid Cymru

While the Plaid Cymru manifesto acknowledges (as do Labour and the Greens) that the Tories have damaged workers’ rights in office, the party’s manifesto has no mention of AI or the effect of digital technology on work. This is surprising given AI is a major focus of both the Welsh Government’s 2019 report ‘Wales 4.0: Delivering Economic Transformation for a Better Future of Work’ and concerns raised by TUC Cymru earlier in 2024.

Reform

Nigel Farage’s party promises a ‘British Bill of Rights’. Largely framed as a measure to guard against the prospect of future pandemic-style lockdowns, it also promises to guarantee the protection of citizens’ data and privacy. Its ‘contract’ states: “Surveillance of the public must be limited and those monitoring us held to account.”

Workers Party

George Galloway’s party intends to observe artificial intelligence “with great attention” but says it is too early to have a specific policy given the potential scale of the technology’s implications. Nonetheless the party promises to ensure AI operates in the interests of the working class and is under “sufficient community control”, bucking the trend of prior industrial revolutions.

Alliance

Alliance’s manifesto promises to demand ‘better regulation’ of AI and social media, and says the technologies represent forces for good and danger depending on how they are harnessed. The party says AI’s implications are widespread, affecting democracy, work, the arts, the economy and citizens’ safety. It is also concerned about the use of machine learning algorithms on social media, and holds little hope of self-regulation within the sector.

Sinn Féin

Sinn Féin’s manifesto does not mention AI or the digital transformation of work.

Note: The Social Democratic Labour Party and Democratic Unionist Party manifestos were not available at the time of writing.

Conclusion


It is notable that the final report of the ‘governance of artificial intelligence’ inquiry only mentions challenges to jobs, work and workers in passing – especially at a time when, as we know, the employment context is one of labour squeeze, low quality jobs, and the rise of in-work poverty. Meanwhile the ‘case’ for AI is taking form at a time when the government – both outgoing and incoming – is scrambling to solve a national productivity crisis and hunting less for magic money trees than a magic growth mist.

Meanwhile, the truth remains that trade unions are primarily looking to legislative protection for safeguards around the future of work. There is no silver bullet – and, as some manifestos acknowledge, there is an added regulatory challenge because AI isn’t emerging in one single place. While the desire for a single regulatory locus of framework is understandable, workers and the public need to be understood as being disproportionately affected by the present and forthcoming risks of AI, and such effects not simply a caveat to the wishes of business. Our data is ultimately the product, after all – which is why we need to be able to cut through the hype and have an honest conversation about what artificial intelligence holds for us all.


Craig Gent is a researcher, writer and editor based in Leeds, and the author of Cyberboss: The Rise of Algorithmic Management and the New Struggle for Control at Work. His research focuses on the politics of algorithmic management, automation and AI from the perspective of the workplace, and he has a growing interest in workplace approaches to AI ethics. He is also a director at Novara Media, where he has worked for over ten years. Craig was previously a postdoctoral fellow at the University of Warwick.