Policy Brief to the Economy and Fair Work Committee of the Scottish Parliament
Policy Brief to the Economy and Fair Work Committee of the Scottish Parliament
Automated discrimination in employment processes: Why Scotland should supervise automated employment algorithms.
“When no one cares you eventually cause a disturbance”
Guardian 2011
Link to PDF version
Executive Summary
There can be no doubt that Artificial Intelligence (AI) is influencing the recruitment process. While there are advantages to the use of this novel technology, with the possibilities of efficiently streamlining the recruitment process, these benefits are juxtaposed against concerns over bias and discrimination, especially towards older workers.
Reports show that older job seekers often face discrimination due to algorithms used in the recruitment process resulting in potential bias and age discrimination. It has been demonstrated that jobseekers in the over-50 age cohort are often experiencing Ageism when attempting to gain paid employment.
With an aging population, this problem will continue to grow and needs to be addressed. This brief outlines both European and International Human Rights policy that has made in-roads into addressing this issue and outlines potential Scottish Governmental alignment in this respect.
It is proposed that the Scottish Parliament’s Economy and Fair Work Committee should therefore:
– Analyse Labour Market statistics to gauge how the use of AI in the recruitment process affects the labour market in Scotland
– Consider the existing EU AI Act and how Scotland could benefit from alignment.
– Consult with the new Scottish Government Commissioner for Older People regarding the use of AI in the recruitment process.
Introduction
“The Scottish Parliament’s Economy and Fair Work Committee is responsible for the supervision of Labour market policy, Equality in the Employment Sector and Fair Economic Development” (gov).
It follows then, that over-seeing the use and regulation of AI technologies in the recruitment process clearly falls under the remit of the Committee.
The use of Artificial Intelligence (AI) in the recruitment process is uncovering economic and employment rights questions (Roumbanis,2025). While AI has the potential to optimise the recruitment process and reduce the cost for companies (Atkinson, Wright 2019), studies and the Equality and Human Rights Commission highlight that older participants, among other groups, are systematically disadvantaged (Chen, 2023; Stypinska,2022; EHRC). The many biases inherent in Automated Decision Making are well documented (Adams-Prassl. Binns. Kelly-Lyth 2022). Ageism is one of these evident biases, impacting negatively on employment chances for certain demographics. Nevertheless, it has been shown that many in the over 50 age group still need and/or want to work (Figure 1). There is an obvious disparity that needs addressing.
Background and Legal Context
Longstanding evidence demonstrates that unemployment has negative consequences on health and mental health (Linn, Sandifer, Stein 1985; Yang, Niu, Amin, Yasin 2024) and can affect the young (Thern, de Munter, Hemmingsson 2017) and the older generations (Kang, Kim 2022). With demographics in Scotland demonstrating a heavy age pyramid, with a disproportionate distribution of older people (National Records of Scotland,2023) coupled with demonstratably Ageist (and sexist) Tech sector (Diginomica;CWJobs), it is clear that the AI recruitment process needs special attention. This is crucially significant as the state pension age rises (ILC 2024) and Ageism comes with a sustained economic cost. For example in the USA Ageism led to excess costs of US$63 billion (WHO 2021). It is clear that no country can afford to maintain discriminatory ageist recruitment practices.
In a positive move, the Scottish Government has formulated ethical principles for the use of AI (Scottish Government, 2021), also creating a voluntary AI register (gov.scot) to help transparency. Nevertheless, there is a lack of clear legal regulation for algorithmic decision making processes used in the recruitment sector in Scotland. The EU has already taken action with the AI Act (EU 2023), while Scotland has seemingly not yet implemented similar safeguards. The Scottish Government has aligned with the EU (Constitution, Europe, External Affairs and Culture Committee) and has also international human rights alignment in the pipeline (planned Human Rights Bill Scotland 2024) but this has hit a delay through the constraints of the devolution settlement (Sommerville 2024, Supreme Court UK 2021).
The Human Rights Law Perspective
International
The right to be protected from discrimination and to be equal in front of the law is enshrined in Article 7 of the Universal Declaration of Human Rights (UDHR) and article (3)26 of the International Covenant on Civil and Political Rights (ICCPR)
The International Covenant on Economic, Social and Cultural Rights (ICESCR) and the ICCPR are part of International bill of Human Rights. In Article (3)7 of the ICESCR, accepted and ratified by the UK, the right to enjoy work without discrimination is accepted by all signatories.
It is worth noting that it is not mentioned whether discrimination in the recruitment process is specifically included within these articles.
The UN Guiding Principles on Business and Human Rights (UNGPs) provide some guidance on ethical principles that should be maintained in the growing new technology sector.
Europe
ECHR
Article 14 of the European Convention on Human Rights (ECHR) protects only against discrimination in connection with other rights guaranteed by the Convention and the UK did not sign the Protocol No. 12 to the Convention for the Protection of Human Rights and Fundamental Freedoms (European Convention of Human Rights), which would include companies into the framework. However, there seems to be no right to work, and as such there is very weak to no protection of the recruitment process in this framework and especially not for an algorithmic recruitment process.
The ECTHR has recognised that age is another status in Article 14 (Guide Article 14). The Court has heard cases relating to Agism (Carvalho Pinto de Sousa Morais v Portugal (Schwizgebel vs Switzerland), and employment (Thlimmenos v. Greece) and agreed to an age limit in recruitment in the police in (Ferrero Quintana v. Spain). Alongside this, positive action (a process to eliminate bias) would be allowed under Article 14 (ECHR).
An international Cyber Security and Resilience Convention could protect applicants from international hostile interference into algorithm recruitment (Tarran 2018;Etudo, Whyte, Yoon, Yaragi 2023; Sukumar, Basu 2024) In addition, Age has now become one of the best predictors of political leaning (Fletscher 2021;Figure 2). Analysis shows that anti-democratic extremism is most strongly affected by the male unemployment rate and the white unemployment rate, consistent with the observation that most members of these extremist groups are white men (Crost 2021). This affects the young (Dunne 2015;Siedler 2011) and those experiencing economic crisis if the World Uncertainty index is high (Gozgor 2021). As voting for political parties is age dependant, a cyber attack on
recruitment algorithms could increase division in societies (Statista 2025).
Therefore, there is a high risk that biased algorithmic decision making could affect the self proclaimed goal of the Scottish Government to deliver fair AI.
The upcoming Commissioner for Older People and planned Human Rights Bill Scotland 2024 would certainly help to alleviate the pressure these algorithms will put on older people.
Suggested actions
The Scottish Parliament’s Economy and Fair Work Committee should therefore:
1.) Analyse Labour Market statistics for influences of AI based automated decision making in the recruitment sector and deliver a work force representative of the age composition of Scotland.
2.) The Committee should consider alignment with the EU AI Act 2023 to measure and influence the AI influence over the employment and as a result, recruitment sector in Scotland.
3.) Ask for input from the Commissioner for Older People. Will a remit of this Commissioner to supervise the AI recruitment and if so, what steps will be taken to address the biases noted in this brief.
Conclusion
While bias in automated decision making has been recognised and is evidenced here, there seems to be very little progress at addressing these issues, especially in relation to Ageism. An unhappy sick workforce should not only desired due to economic reasons (ie should not be a burden on the Health and Care Sector) but out of ethical concerns as well. At present, it would appear the algorithms in present us fall well short of the targets that the Scottish Government has set regarding AI technology.
As this brief has pointed out, the demographic structure of Scottish society and the top-heavy population pyramid mean that is essential to address Ageism faced by a growing cohort of the population that can still be economically active.
Additional Reading
- AI ageism: a critical roadmap for studying age discrimination and exclusion in digitalized societies (AI & Society) 2022
https://link.springer.com/article/10.1007/s00146-022-01553-5 - Age-related bias and artificial intelligence: a scoping review (Nature) 2023
https://www.nature.com/articles/s41599-023-01999-y - Ageism in artificial intelligence for health (WHO) 2022
https://www.who.int/publications/i/item/9789240040793
