HR Horizons: Winter 2024

Background

Artificial Intelligence (AI) is not a new concept: it was first developed in the 1950s and is now a regular feature of everyday life – for example, AI is the technology which underpins predictive text in messaging apps, image recognition and customer service ‘chatbots’. Broadly, AI is defined as machines which might be made to work in the same way as humans. The launch of generative AI based on Large Language Models (LLMs) at the end of 2022 marked a step change in the development of AI technology: these models use deep learning from ‘scraping’ the entire internet to produce natural language texts based on information requested in the input (known as a ‘prompt’). So, for example, an LLM app such as ChatGPT can produce comprehensive information on almost any subject or question within a few seconds; generative AI can also produce images and logos, write and correct computer code, create spreadsheets, summarise complex topics and design campaigns and strategies.

But AI/LLM technologies do come with some significant health warnings: there are numerous issues with data security, accountability and transparency; and LLMs are prone to ‘hallucinations’: responses that read as if they are accurate but are produced when an LLM scrapes the internet for data but cannot find the information it needs (a recent First Tier Tax Tribunal hearing for which AI produced nine ‘helpful’ but fictional legal precedents was recently reported in the Law Society Gazette).

What’s next?

Many commentators are convinced that AI has the potential to be a powerful ‘co-pilot’ tool for productivity: a 2023 study published by Stanford University and the Massachusetts Institute of Technology found that the use of AI at one tech company increased productivity by 14%. Reportedly, the CEO of JP Morgan has said that thousands of JP Morgan staff are already using AI at work and this will eventually allow the company to operate a 3.5 day working week as standard. This is echoed by an academic at the London School of Economics who anticipates that AI is the gateway to normalising a universal reduced working week. So, looking ahead, it seems very likely that the expansion of AI will trigger shifts in working arrangements and possibly wholesale changes to the labour market, with generative AI impacting more on skilled and creative roles compared to ‘conventional’ AI: according to a November 2023 report by the Department for Education , roles which require a higher level of education are more likely to be affected by AI and LLMs, with those most affected being those in London and the South East of England and workers in professional occupations (law, finance, business management) and insurance.

There is no specific ‘AI legislation’ in the pipeline in the UK. The government’s August 2023 White Paper, A Pro-innovation Approach to AI Regulation confirms that it intends to take a ‘light touch’ approach to AI regulation. This will be based around principles and guidance rather than legislation because the legislative process would be too slow to keep pace with the rate of change. Regulators such as the Information Commissioner’s Office, the Equality and Human Rights Commission and the Health and Safety Executive will be responsible for producing AI guidance for their specific areas of responsibility. The government will publish a Regulatory Road Map in March 2024 and will consult on the creation of an AI risk register to support regulators’ internal risk assessments.

The EU is taking a more prescriptive approach than the UK, with a draft EU AI Act being prepared by the European Commission which will apply to any AI high-impact general purpose AI models which affect people located in the EU (regardless of whether the system is provided by an organisation inside or outside the EU). The European Council’s December 2023 press release on the new Act is here.

Things to consider

The first and most important decision for senior leaders will be whether there is a use case for AI technology; rather than rushing to embrace a new technology for its own sake. Key areas in which AI might assist include recruitment (candidate sourcing and screening), automating business processes, ‘reading’ and summarising complex documents, data analytics and forecasting. But employers will need to understand what types of technologies best perform which tasks, and the strengths and limitations of each.

Although there is no specific AI employment legislation on its way in the UK, employers must be alive to the fact that they will still be required to comply with their usual obligations under the existing employment law framework, notwithstanding that tasks and / or decisions have been outsourced to AI. The complex nature of the technology can make it more difficult for employers to explain or justify decisions that have been made with the assistance of AI. And the inability of AI technology to understand data in context or to apply critical thinking to raw data may not always sit well with some of the very human concepts which underpin the employment law framework – for example, ‘fairness’, ‘reasonableness’ and ‘unfavourable treatment’. It will, therefore, be important to ensure that there is some human critical analysis of AI processes and that there is transparency in AI decision making. AI systems are vulnerable to biases which have been ‘baked in’ the system: whether because the people who have built the systems have incorporated their own biases when creating the technology or whether because there is bias present in historical data. Regular audits and assessments of AI systems will help to identify and reduce the risk of any inadvertent adverse impacts.

Any organisation utilising AI will, therefore, need to think about how AI is understood across the workforce, ensuring that senior leaders and managers have a solid grasp of how AI technology is applied and, importantly, its limitations and risks. Working with IT teams at an early stage will be important, and ensuring that governance frameworks are drafted at an early stage; these will include scoping the use of AI, agreeing limitations on use, data protection/privacy policies, and introducing an AI policy for employees (to include scope, ethics, training and enforcement).

In summary, six key areas of AI/LLM focus for HR leaders and In-house Counsel will be:

1. assessment of the potential benefits/risks of AI

2. data security and privacy

3. governance (AI policies and overlap with other policies and procedures)

4. training

5. workforce communication and education, and

6. diversity and inclusion, transparency, and fairness.

Organisations with operations in the European Union will need to track the progress and the EU AI Act and ensure compliance if utilising AI systems within the EU.

Please see our Insight, ‘Harnessing Technology’ for tips on what to think about before drafting an AI policy and listen to our AI podcast for further insights on employment law and data protection risks around AI at work and how to deal with them.

"AI is coming for all of us! Whilst the practical application is still a work in progress in many contexts, this looks like changing quickly not least as a result of potential cost savings. In the early days, some careful governance looks like an essential step to ensure companies maintain control of internal application and can take a considered view and apply appropriate risk mitigation."

Sarah Skeen
Partner

Background

We reported in the last edition of this report that employers should expect greater scrutiny of compliance with labour rights, such as holiday pay, minimum wages, statutory sick pay and prevention of modern slavery, ahead of the creation of a new Single Labour Market Enforcement Body. Momentum on this proposal has now stalled. The government’s current position is that it is reviewing what the creation of a new Enforcement Body would mean in the light of limits on parliamentary time (see page 18, House of Commons Business and Trade Committee report, July 2023).

We flagged in our last update that the government had published a Future of Work Review, to determine labour market policy and strategy (including the government’s response to the Taylor Review on Good Work). There is no significant progress to report on the Future of Work Review, except that the MP leading the review has recommended that the government focuses on four areas in greater detail: AI and automation, skills, place and flexibility, and workers' rights.

What’s next?

The government proposes to meet some of the objectives of a Single Enforcement body through its 2023/2024 labour market enforcement strategy (published on 23 October 2023). The government’s strategy states that it is considering how to improve joined up monitoring, thinking and communication between existing enforcement bodies, such as HMRC and the Employment Agency Standards Inspectorate.

Should the Labour Party win the General Election due to take place by January 2025, employers can expect significant changes to the employment law landscape. These have been set out by the Labour Party in their Green Paper, New Deal for Working People, in the speech of the Shadow Deputy Prime Minister (Angela Rayner MP) to the 2023 Labour Party Conference and reports in the press. The Labour Party have pledged to introduce an Employment Bill within the first 100 days of entering office.

Key employment reforms proposed by the Labour Party include the following.

• A standard six-month time limit for bringing all Employment Tribunal claims (currently usually three months)
• Removal of statutory caps on compensation for unfair dismissal.
• Removal of the two-year qualifying period of employment for unfair dismissal claims, so that the right to be fairly dismissed becomes a ‘day one’ right.
• Removal of minimum service levels (MSLs) during strikes. Labour would repeal the Strikes (Minimum Service Levels) Act 2023, which provides for MSLs in specific services such as health, transport and education.
• The introduction of electronic balloting and bolstering trade union rights to access within workplaces. The Labour Party has also pledge to “boost” collective bargaining, starting with a Fair Pay Agreement in adult social care.
• Ending ‘fire and re-hire’ and extending information and consultation procedures so that employers are required to reach agreement about contractual changes with their workforce.
• Introducing a new ‘right to disconnect’. This would introduce a new statutory right to be unavailable for work outside of specified hours. There would also likely be linked protections from harassment and victimisation for exercising the same right.
• The Labour Party’s Green Paper originally pledged to create a single status of ‘worker’ (rather than existing categories of ‘employed/self-employed/worker) but it has been reported in the press that the Labour Party is now rowing back from that and instead will introduce two categories of employment: workers and the genuinely self-employed.
• Create a single enforcement body to enforce workplace rights (similarly to the body proposed by the existing government – see above).
• Ban zero-hours contracts.
• Introduce mandatory disability pay gap reporting for larger employers
• Require organisations with more than 250 employees to introduce menopause action plans.
• Possible re-introduction of the obligation to protect employees from third-party harassment and extending the forthcoming obligation to take reasonable steps to prevent sexual harassment to take all reasonable steps to prevent harassment (as originally proposed by the current government).

Things to consider

Although legislation creating a standalone enforcement body has been put on hold for now, and seems unlikely to see legislative light of day before the next General Election, employers are unlikely to entirely avoid increased scrutiny of compliance with employment rights. Government policy has tended to focus on greater compliance and enforcement of existing obligations to staff rather than introducing new legal requirements.

Employers must, therefore, remained focused on audit and compliance given the cost and reputational risks of being investigated by an enforcement body such as HMRC on National Minimum Wage or the Information Commissioner’s Office on data security.

Given that current opinion polls are pointing strongly to a Labour Party victory in the next General Election, it would be prudent to expect significant employment law reform in the medium-term future. The extension of the three-month time limit for bringing Employment Tribunal claims would immediately open employers to increased risk of claims, with the potential cost of claims increasing with the removal of the statutory cap of one years’ wages for unfair dismissal claims. However, it is the removal of the two-year qualifying period for claims of unfair dismissal which would probably have the most significant impact for employers: if passed, it would mean that from day one of employment, all employees would be entitled to fairly dismissed. This would involve significantly increased legal risk, cost and management time for all employers.

The proposed introduction of a legal right to disconnect echoes similar protections in a range of other countries. For more information on what this might mean for employers if enacted, please see our recent article in People Management magazine.

“It is worth remembering that governments of all shades have for decades used employment/labour law reform as a relatively cheap way of attempting to send a message as to their political priorities and values- think the trade union reforms of the 1980s, the introduction of the national minimum wage and tribunal fees. Although we can expect some rowing back from pre-election pledges, a substantial swing in favour of enhanced individual rights and possibly increased trade union influence, is looking increasingly likely.”

Amy Stokes
Partner