Employment and immigration | UK Regulatory Outlook November 2022

Are new employment laws on the cards?

In our recent Insight we outlined that a number of bills have been pushed forward in light of recent industrial action with the aim "to ensure more people stay in work". These include:

• a new bill to create a framework enabling the government to implement minimum levels of service to be provided by "specified transport services" during strike action;
• a private member's bill that will entitle an employee to take one week of unpaid leave a year to provide or arrange care for another; and
• a private member's bill offering pregnant women and new parents greater protection against redundancy.

Of particular importance, regulations extending the ban on exclusivity clauses in employment contracts to low-income workers earning no more than the lower earnings limit (currently £123 a week) have been made and will come into force on 5 December 2022.

Exclusivity terms are defined as any contractual term which prohibits a worker from doing work or performing services under another contract or arrangement, or which prohibits a worker from doing so without their employer's consent.

Workers are protected from detriment if they breach an exclusivity term in their contract and employees benefit from automatic unfair dismissal protection (with no qualifying service requirement) if the reason or principal reason for the dismissal is that they breached an exclusivity term.

ICO guidance on how to use AI and personal data appropriately and lawfully

As noted in our previous Insight, we have been awaiting guidance from the UK Information Commissioner's Office (ICO) with regard to employers implementing artificial intelligence (AI) recruitment software.

The ICO have now issued new guidance: "How to use AI and personal data appropriately and lawfully" that will be very relevant to recruiters and people platforms who are looking increasingly at how they can improve services with the use of AI.

As well as advice to businesses, the guide is also intended to provide a grounding for regulators (including no doubt the Employment Agencies Standards Inspectorate) assessing the use of AI in the sector(s) they regulate. The guidance is comprised of a set of useful tips on how to improve the use of AI, plus some frequently asked questions on AI and personal data. In its top tips, it suggests:

• taking a risk-based approach when developing and deploying AI;
• thoroughly considering how to explain the decisions made by the AI system to individuals affected;
• collecting only the data which is needed to develop AI system;
• addressing risks of bias and discrimination at an early stage. This will be key for recruiters given the level of liability they may suffer if they are involved in discrimination. The guidance says that "training data for an AI system designed to assist in recruitment decisions may reflect discrimination where men have historically been considered to be more suitable candidates for certain roles. You could either modify the data, change the learning process, or modify the model after training. There could be bias in the way variables are measured, labelled or aggregated, or developers could demonstrate biased cultural assumptions. If this is due to human labellers, you should ensure there are clear criteria and lines of accountability";
• taking time and dedicating resources to prepare the data appropriately;
• making sure that the AI system is secure, and that any human review of decisions made by AI is meaningful; and
• engaging an external supplier to ensure that the use of AI is appropriate.

Addressing risks of bias and discrimination at an early stage will be key in particular for recruiters given the level of liability they may suffer if they are involved in discrimination. The guidance says that "training data for an AI system designed to assist in recruitment decisions may reflect discrimination where men have historically been considered to be more suitable candidates for certain roles. You could either modify the data, change the learning process, or modify the model after training. There could be bias in the way variables are measured, labelled or aggregated, or developers could demonstrate biased cultural assumptions. If this is due to human labellers, you should ensure there are clear criteria and lines of accountability".

Blackouts this winter? What employers need to consider

With warnings from the National Grid of the potential for blackouts this winter, in our latest Coffee Break podcast, Danielle Kingdon, partner in our employment team, looks at the legal issues for employers and the practical steps to consider. This is an issue on which we expect to see further developments; if you would like to discuss the issues further, please do contact your usual Osborne Clarke contact, Danielle or Mary Lawrence, partner in our Health and Safety team. 

ICO launches consultation for draft guidance on workers' health information

On 27 October 2022, the ICO published draft guidance on handling workers' health information for public consultation.  The aim of the draft guidance is to provide practical insights for employers, and promote good practice, on how they can handle the health information of their workers in accordance with data protection legislation.  The consultation forms part of the ICO's wider efforts to replace its employment code of practice with new, UK GDPR-focused guidance.