Artificial intelligence is revolutionising the workplace, significantly reshaping how organisations recruit, manage, and promote talent. In Australia, for example, this transformation is particularly pronounced in recruitment practices, where AI tools are being employed at an accelerating pace.

According to the latest Responsible AI Index, 62 % of Australian organisations reported using AI in recruitment either moderately or extensively in 2024.

These AI systems are designed to streamline hiring by classifying, ranking, and scoring candidates based on attributes such as personality traits, behaviour, and skill sets. They either make decisions autonomously or assist recruiters in determining which applicants progress to subsequent stages.

While these innovations offer undeniable efficiency benefits, they also introduce unique and serious risks of bias and discrimination that must be addressed by HR professionals and business leaders.

Amplifying bias and discrimination

A critical concern is that AI systems operate at speeds and scales beyond human capability, yet their decision-making processes are often opaque. Candidates may be unaware they are being evaluated by AI, and the criteria used can be difficult to interpret or challenge.

Research conducted on this issue highlights that AI recruitment tools can disproportionately disadvantage specific groups, including women, older workers, people with disabilities, and those who speak English with an accent. The use of AI for screening CVs, conducting assessments, or video interviews can inadvertently embed or amplify discrimination. This bias might stem from the training data, the algorithmic design, or even the way an organisation implements the technology.

For example, many AI screening platforms are not validated for accessibility, excluding candidates with disabilities from fair evaluation. One career coach interviewed during the study shared a striking case of a neurodivergent client – a university top-performer – who consistently failed personality assessments because the AI interpreted his atypical responses unfavourably. This individual’s promising credentials were overlooked due to an algorithm’s inability to accommodate neurodiversity.

The problem of transparency

A further challenge relates to transparency and candidate experience. Time constraints for answering AI-generated interview questions are often inadequately communicated, creating stress and unfair disadvantage. One career coach noted instances where candidates were abruptly cut off mid-response due to unclear time limits.

More broadly, job seekers – especially those with disabilities – often face a lack of information about the recruitment process. Without transparency, candidates cannot effectively advocate for themselves or request necessary accommodations, exacerbating existing inequalities.

New barriers to employment

AI recruitment systems also introduce structural hurdles beyond bias. Candidates require access to technology, including a reliable phone and stable internet connection, along with a baseline of digital literacy, just to participate in AI-driven assessments. This requirement disproportionately impacts marginalised communities and those with fewer resources.

Such barriers can discourage applicants from applying or cause them to withdraw midway through the hiring process, further narrowing the talent pool and undermining diversity and inclusion efforts.

Legal and regulatory lag

While AI is advancing rapidly in HR, legal frameworks have yet to catch up. Current employment laws and anti-discrimination regulations do not fully address the nuances introduced by AI recruitment technologies, leaving many employers vulnerable to legal and ethical risks.

This means adopting AI tools cautiously, ensuring fairness and transparency, and actively monitoring for unintended consequences. Companies must consider validating their AI systems for accessibility and inclusivity, providing clear communication to candidates, and maintaining a human element in recruitment decisions.