HR remains cautious around AI implementation
New research by Winmark showed just…
…of HR functions surveyed were using AI
Lower than the all-department average of...
... half of which were at a mature stage of adoption
However, a further…
... were considering implementation
New research published the global C-suite network facilitator Winmark evidences that HR has one of the lowest levels of adoption of AI across the corporate functions. In its AI Applications Pulse Survey, conducted in August and September 2023, Winmark identified that just 15% of HR functions surveyed were already using AI – notably lower than the 26% all-department average – half of which were at a mature stage of adoption whilst the other half were at an early stage. However, a further 69% were considering implementation.
AI implementation and usage
Source: Winmark, AI Applications Pulse Survey, August-September 2023
Limitations to implementing AI
Source: Winmark, AI Applications Pulse Survey, August-September 2023
When it comes to the challenges of adopting AI, HR departments face some specific hurdles. They often mention not having enough time (31% in HR compared to 27% overall) and lacking sufficient data (31% in HR versus 26% overall). However, they seem less worried about issues like data privacy and regulation (31% in HR versus 48% overall), deciding which projects to invest in (15% in HR versus 27% overall), or getting enough budget (8% in HR versus 17% overall). So, HR departments seem ready to embrace AI if they can access the right knowledge and skills, which is the top barrier cited by 62% of respondents.
And within this top-rated issue limiting progress, there will undoubtably be concern around having the ‘knowledge’ of how to appropriately implement AI in the UK without a specific legislative governing framework. As Winmark notes, “managing the regulatory aspects of internal and external AI operations will be a major challenge for business leaders.”
Furthermore, HR professionals, especially those in multinational companies, are likely familiar with instances in other parts of the world where rules are being put in place to regulate the use of AI in hiring. These rules mainly emphasise transparency, ensuring that the way AI systems are used is clear and understandable, and that their decision-making processes are explainable.
A case in point is the legislation introduced in New York state in July 2023, which requires organisations using software that relies on machine learning or AI to help them select preferred or weed out ‘less-suitable’ candidates (called an automatic employment decision tool (AEDT)) to pass a third-party audit that shows that it is free of sexist or racist bias – and publish their results.2 As, at the latter stages of the drafting, the law was changed to include only those AEDT without any human oversight, very few companies appear to believed that it is applicable to them, instead, likely attesting that there is human involvement somewhere in the process.
2. SHRM, New York AI Law is Bust, 19 Feb 2024: https://www.shrm.org/topics-tools/news/technology/new-york-city-ai-law#:~:text=The%20law%20requires%20employers%20that,candidates%20that%20such%20tools%20are
That said, lawmakers are watching to see what the efficacy of the new law is – and many legal commentators believe that, if there is evidence that people have been harmed through the application of AI in recruitment, there will be further laws passed during the course of 2024 to strengthen its application.
The one to watch, however, is the EU’s AI Plan – “the first comprehensive regulation on AI by a major regulator anywhere.” 3 Of note “high risk applications, such as a CV-scanning tool that ranks job applications are subject to specific legal requirements” - compliance with which will be within 24 months of when the Act comes into force (pending).
3. The EU Artificial Intelligence Act: https://artificialintelligenceact.eu/
Key takeaways and recommendations:
The EU AI Act assigns applications of AI to three risk categories:
- First, applications and systems that create an unacceptable risk, such as government-run social scoring of the type used in China, are banned.
- Second, high-risk applications, such as a CV-scanning tool that ranks job applicants, are subject to specific legal requirements.
- Lastly, applications not explicitly banned or listed as high-risk are largely left unregulated.
The EU AI Act assigns applications of AI to three risk categories:
- First, applications and systems that create an unacceptable risk, such as government-run social scoring of the type used in China, are banned.
- Second, high-risk applications, such as a CV-scanning tool that ranks job applicants, are subject to specific legal requirements.
- Lastly, applications not explicitly banned or listed as high-risk are largely left unregulated.
In contrast, the only activity to date by the UK government has been the publication of guidance, the most recent of which is Responsible AI in Recruitment: Guidance published by the Department for Science, Innovation & Technology in March 2024.
“This guidance identifies potential ethical risks of using AI in recruitment and hiring processes. It further outlines how AI assurance mechanisms can provide organisations with the tools, processes and metrics to evaluate the performance of AI Systems, manage risks, and ensure compliance with statutory and regulatory requirements. It is intended for organisations seeking to procure and deploy AI systems in their recruitment processes.”
Department for Science, Innovation & Technology, Responsible AI in Recruitment: Guidance
“This guidance identifies potential ethical risks of using AI in recruitment and hiring processes. It further outlines how AI assurance mechanisms can provide organisations with the tools, processes and metrics to evaluate the performance of AI Systems, manage risks, and ensure compliance with statutory and regulatory requirements. It is intended for organisations seeking to procure and deploy AI systems in their recruitment processes.”
Department for Science, Innovation & Technology, Responsible AI in Recruitment: Guidance
In the absence of any formal regulatory steer within the UK – and given the clear need for HR to innovate with caution, given the legal, cultural and business implications of doing so - it is unsurprising that HR functions are taking a measured approach to adoption, notably within recruitment. This is reinforced by the fact that just 2% of companies admitted to utilising AI in recruitment within the Winmark August-September 2023 survey. Use cases, to, date, include utilising AI for basic work around job descriptions and LinkedIn profile analysis. The use of AI in recruitment and candidate sifting is, however, firmly on the planned list for near-term adoption.
In the absence of any formal regulatory steer within the UK – and given the clear need for HR to innovate with caution, given the legal, cultural and business implications of doing so - it is unsurprising that HR functions are taking a measured approach to adoption, notably within recruitment. This is reinforced by the fact that just 2% of companies admitted to utilising AI in recruitment within the Winmark August-September 2023 survey. Use cases, to, date, include utilising AI for basic work around job descriptions and LinkedIn profile analysis. The use of AI in recruitment and candidate sifting is, however, firmly on the planned list for near-term adoption.
As to how aware of employer AI practices candidates currently are – and, crucially, how they feel about it - recent research by recruitment technology provider Criteria shows that around a third (29%) are very aware of hirer AI-assisted recruitment tactics and, on balance, slightly more are positive (26%) about its use than are negative (22%).5
5. Criteria, Candidate Experience Report 2024
How well do you understand how employers are using AI or other technology to hire?
29% Very well
45% Somewhat well
21% Not very well
5% Not at all well
How do you feel about the idea of employers using AI during the hiring process?
26% Positive
51% Neutral
22% Negative