AI in hiring: bias & privacy an issue for 40% of hiring teams
Workable's AI in Hiring & Work survey finds that hiring bias is a major issue for 40% of hiring team managers when using AI in hiring. Privacy and compliance are also top concerns according to the survey. This is an excerpt from our AI in Hiring and Work survey report, based on responses from 950 hiring managers in the US and UK. Visit here to download the report in full.
There are many positives when using AI technology in the hiring process – but let’s be real: there are drawbacks as well.
So, in our AI in Hiring & Work survey, we asked respondents what issues they’ve run into when using AI in hiring. What struck us is how spread out the responses were in comparison to other questions asked in the survey.
That being said, however, we know the concerns around bias and privacy when using AI at work – and we’ve written extensively about both, including best practices for maintaining ethical AI usage, overcoming bias using ChatGPT, and tackling bias and privacy concerns when using AI in HR.
So it came as no surprise that bias and privacy were leading concerns in our survey dataset. Let’s go through them now.
Hiring bias
In the survey, two out of five (40%) of respondents pointed to hiring bias as a major issue when using AI in their recruitment.
Fair enough. Bias is a huge consideration when making a hiring decision in general. Technology does help in overcoming bias in some areas with anonymized screening, standardized assessments, and other features.
And it’s important to note that AI tools are often trained on existing materials and experiences, meaning it’ll aim to replicate the biases inherent in the system. So, as we’ve previously reported – AI is not at fault. Rather, the data it’s trained on is.
In this case, technology giveth and it also taketh away. However, hiring teams can overcome this challenge with the right level of human involvement and supervision, if not taking over outright.
Legal considerations
Meanwhile, 37.2% point to privacy concerns especially when handling the personal data of candidates and employees.
Perhaps overlapping is the 30.7% who highlight compliance as a focal area that’s keeping them up at night – largely due to copyright, security, and other regulated areas.
Data privacy, of course, is a major ongoing concern for employers with mounting legislation on the heels of GDPR in Europe and CCPA in California. It’s reaching a point where every government will have some form of data privacy law in place.
The big concern with AI is that hiring teams interact with external technology using sensitive candidate and employee data in what amounts to a new wild west in this age of AI. Legislation around this is sparse, although is rapidly evolving over the coming years – we’re already seeing this in the European Union’s AI Act. There’ll no doubt be more to come.
Meanwhile, an AI tool usage policy will be useful for your organization.
Talent identification
In other parts of the survey, we found the top two use cases of AI in hiring to fall in line with candidate identification – resume screening tools being used by nearly three of five (58.9%) and candidate matching tech being utilized by 43.1%.
We find that the major issues are in line with those top use cases. For instance, overemphasis on keywords (31.2%), inaccurate interpretation of soft skills (26.3%), inability to capture candidate potential (15.5%), and over-reliance on historical data (15.5%) are all popular areas of concern for hiring team members when using AI.
Despite its rapid evolution, AI and the many tools utilizing it continue to be in a relatively nascent stage. Two scenarios are likely here:
AI tools are not quite sophisticated enough to support teams in identifying top talent
Hiring team members are not quite sophisticated enough in how they’re using AI tools
It’s probably a combination of both.
The industry lens
Hiring bias is an even bigger consideration for Construction (48%), Manufacturing (45.5%) and IT / Technology / SaaS (44.5%) versus 40% overall. It’s not nearly so much of a concern for Accounting (35%).
Data privacy, on the other hand, is a major concern for Accounting / Finance (43.8% vs. 37.2% overall) and for IT / Technology / SaaS (44.9%). It’s not as high in the minds of Healthcare (29.2%) or Retail (32.8%).
Likewise, compliance is top of mind for Accounting / Finance (43.8% – 13.1 full points higher than the overall 30.7%) and IT / Technology / SaaS (39.2%), but not so much for Construction (22.7%) and Retail (22.4%).
Healthcare, meanwhile, lamented the overemphasis on keywords (41.2% vs. 31.2% overall) and Construction downplayed that impact on their own hiring processes (22.7%).
Frequently asked questions
- What are the main concerns with using AI in hiring?
- The survey highlighted hiring bias (40%), data privacy (37.2%), and compliance issues (30.7%) as the top concerns. Additionally, overemphasis on keywords (31.2%) and inaccurate interpretation of soft skills (26.3%) were significant, pointing to the complexity of AI's integration into fair hiring practices.
- How does AI contribute to hiring bias?
- AI's contribution to hiring bias stems from its reliance on pre-existing data, which may contain inherent biases. This was a concern for 40% of our respondents, underscoring the need for careful management and oversight to mitigate bias in AI-driven recruitment.
- What are the privacy and compliance risks associated with AI in recruitment?
- Privacy risks, cited by 37.2% of respondents, involve the handling of sensitive candidate information, while compliance risks, noted by 30.7%, relate to legal standards in copyright, security, and other areas. These issues reflect the urgent need for robust policies and practices in AI deployment.
- Can AI tools accurately assess soft skills and candidate potential?
- While AI offers many advantages, 26.3% of survey participants raised concerns about its ability to accurately interpret soft skills, and 15.5% doubted its capacity to assess candidate potential. These figures highlight challenges in developing AI that can comprehensively evaluate candidates beyond technical qualifications.
- How do industry-specific concerns vary regarding AI in hiring?
- Concerns vary significantly by industry: Construction (48%) and IT/Technology/SaaS (44.5%) reported higher concerns about bias. Data privacy was a major concern in Accounting/Finance (43.8%) and IT/Technology/SaaS (44.9%), indicating diverse priorities and challenges across different sectors in adapting AI for hiring.
