Skip to content
ADVERTISEMENT
Featured

As AI reshapes hiring, HR is left holding the risk 

Barb Hyman, CEO of Sapia.ai, discusses how AI-driven recruitment is creating new transparency and bias challenges for HR teams.

5 min read
Share
Barb hyman sapia ai
ADVERTISEMENT

Each week, we see headlines of highly qualified people struggling to secure roles despite sending hundreds of applications. In search of employment, many applicants are now turning to artificial intelligence (AI) to update their CVs and present themselves in the best way possible, or to automate the process of tailoring their credentials to meet job specifications and hit all the key criteria. On the other side of the table, HR teams are also relying on AI systems as the volume of applications has become overwhelming.

AI is becoming increasingly influential in every stage of recruitment, from screening candidates to making the final call on who gets the role. But with AI’s involvement in so many decisions, there is a question over who is accountable for the outcomes?

In most organisations, HR has ended up responsible for the outcomes of AI hiring, as a by-product of adopting these new tools. The result? HR is now accountable for decisions shaped by systems they did not design and cannot fully control.

This is widening the disconnect between accountability and control

HR teams are now expected to use AI in recruiting to speed up the process, yet few truly understand how these systems are executing crucial decisions. These tools can significantly reduce time-to-hire, effectively cutting costs and allowing HR teams to focus on more pressing, strategic tasks. 

The risk? Bias and regulation become hard to navigate when the technology itself is a black box. Candidates are searching for feedback on how their applications are assessed. When something goes wrong, it is HR that takes the hit. Responsibility for AI risk has landed with HR, but without the structure or support needed to manage it.

For example, candidates are being rejected by AI screening tools without any feedback or explanation. They then ask HR teams for feedback on what went wrong, but are often ghosted because HR teams cannot give an accurate response if they themselves do not understand how the AI screening tools reached their decision. This in turn leads to negative candidate experiences, complaints, and reputational damage, all of which fall back on HR teams and leave them exposed to bias risks due to the lack of transparency in the process.

Where do we go from here?

ADVERTISEMENT

Too often, HR is handed new tools with little to no say in how or why they are introduced, but what if we could change that? By involving HR from the offset and asking the right questions about where the technology comes from, how it works, and the value it brings, we could shift from adoption to active ownership.

You don’t need to be an expert on how these systems operate, but understanding the basics, on how the technology works and where its limits lie is crucial. Without this, it’s all too easy for hidden assumptions to shape your decisions.

Picture two candidates applying for similar roles. In one circumstance, an applicant is screened out by AI simply because their CV does not match the keyword patterns the model has learned to prioritise. They receive an automated rejection, with no explanation. Neither the candidate or the HR team realises that the system has quietly started to favour certain career paths, filtering out qualified people without warning or oversight. The outcome? A decision that feels as unpredictable as rolling the dice in a game of Snakes and Ladders where the number rolled might see you climb a ladder or slide down a snake to a stage they’ve already been. 

ADVERTISEMENT

Compare this to organisations that use AI to support, rather than remove, human input. Here, candidates know that AI is part of the process and they are given feedback on how their application was processed. Recruiters are given insights about the candidate that explain the scoring in detail. Such systems are built with guardrails that are regularly reviewed by HR employees, meaning they are built to deliver efficient results without sacrificing fairness. In this case, the experience is accountable, and most importantly, human.

Turning uncertainty into clarity

For candidates in the modern era of hiring, a simple, point blank rejection does not suffice and transparency has become a baseline expectation. Regardless of its early or final stage of the process, if candidates are refused progression, there are increasingly common demands to understand what data has been used, how it was treated and the rationale behind the rejection. This is heightened to new levels of scrutiny when rejection comes before any humans are noticeably involved, such as when AI platforms decide a candidate’s potential based on the credentials listed on application forms, CVs, or worse, data gathered without the candidate’s consent. 

The real question that talent acquisition teams must ask themselves is whether through AI tools or not, they are collecting enough knowledge or data with which to stand behind the decisions that are being made. If an organisation can’t explain how their software works or where their decisions come from, can it really trust the outcomes? At the end of the day, you cannot support your stakeholders if you lack confidence in the systems shaping your decisions and when it comes to hiring, we can’t overlook the human impact it has on jobs, careers, and livelihoods.

Barb Hyman is CEO of Sapia.ai

Related

ADVERTISEMENT