AI recruiting in practice
Early adopters of AI-driven recruiting software (2017) reported the cost per candidate screen reduced by 75% and staff turnover by 35%, indicating the positive effects of the means to process large volumes of data effectively and arrive at insights that make intelligent hiring decisions.
Effects continue to be positive. The recent LinkedIn Global Hiring Trends reported recruiters and hiring managers worldwide agreed AI is a disrupter and is helping in the following key areas:
- Saving time (63% agreed)
- Removing human bias (43%)
- Delivering the best candidate matches (31%)
- Helpful when sourcing candidates (58%)
- Screening (56%)
- Nurturing Candidates (55%)
Research by Korn Ferry backed this up with close to two-thirds (63%) of talent acquisition professionals reporting the recruiting process at their company as much improved since AI was introduced. More than half said candidate selection via algorithm had resulted in a more accurate section for interviews, and roles filled faster.
Further, AI advantages include a more engaging user experience for hirers and candidates; automated high volume administrative tasks; chatbots as an additional channel of engagement; elevated employer brand, and adherence to diversity, inclusion, and equality policies.
Benefits withstanding, there are concerns some AI tools could infringe privacy rights, with AI practices, such as facial recognition, considered unethical and beyond the needs required to appoint the best candidates.
There is still a lot of work to ensure that biases aren’t built into algorithms and that people without digital access to recruitment tools are not excluded from the process.
Further AI concerns include reduced human interaction and lack of human judgment; biases creeping in at all stages from programming to managing algorithm; instances of inaccuracy, especially where data was not accurate; potential PII data breaches, ethical issues, and a negative effect on addressing DI&E policies.
Human bias and recruiting efficiencies
According to the Cognitive Bias Codex, humans are subject to more than 180 cognitive biases in the traditional hiring process. from reviewing a job applicant’s information to selecting for an interview and conducting interviews.
With AI, biases will show up in patterns found in legacy recruitment data. The longer the TA history, the more the AI-powered process can learn, preferences removed, and the function “re-educated.” Several approaches can be followed.
One is to have two algorithms working in parallel to mitigate bias. The first algorithm selects a candidate based on a specific data set, i.e., skills required. The second is responsible for underlying sensitive attributes such as gender, age, race, and school-leaving qualification. The first selects the best candidate based on attributes required for the role. The second seeks out sensitive attributes to ensure no bias in the selection process.
Post-processing is another technique. It’s designed to take inherently biased results and recast them fairly and accurately. Suppose the distribution of job applicants is not equal regarding gender, with a danger one gender might be ranked higher. In that case, the system readjusts, so each application has the same chance to be reviewed and considered.
Pros and cons of AI for an employer brand
AI can make HR processes more efficient. At one time, applications were mailed, and fingers crossed, the postal system delivered. The invitation for an interview, or otherwise, was reliant on the postal system.
Email replaced this and then, more commonly now, the application portal. This is managed by an algorithm, offering rapid candidate response. Job advertisements are optimised for a target group using augmented writing, and the talent search is simplified by matching attributes against applicant algorithms. A common bias here is the expectation the candidate is degree-educated. Without this box checked, years of relevant experience are not registered.
First stage interviews are made more flexible with time-delayed video interviews and assessed with the aid of AI, the suitability of applicants compared with data from already successful employees.
While the benefits of AI in the hiring process are clear for employers, from a candidate perspective, it can be off-putting. The lack of human instance can scare off potential employees, and the oversight of ideal experience is already mentioned. Further, the potential is opened for legal action, regulatory fines (notably from Europe’s General Data Protection Regulation), shareholder and employee concerns, and reputational harm.
Reducing risks in recruiting with AI
To evaluate the associated risks better, it’s necessary to maintain transparency in the system’s decision-making. Some critics say it’s almost impossible to determine how the algorithms arrived at the result.
In terms of recruiting and hiring, this explains why candidate A was selected over candidate B. Explaining the decision could be more accessible when the person in charge of the company knows how the system works in the background. This takes us into another topic of conversation; the new skills of the HRD.