Organizations have been increasingly moving toward automating many of their functions and processes for the last few years to improve their efficiencies, minimize costs, and boost their revenues. The COVID-19 pandemic provided the impetus for automation. The human resources (HR) function is no different. But given that many HR-related issues are complex and need a human element, is artificial intelligence (AI) ready to automate HR-related tasks?
Skynova recently surveyed close to 200 HR professionals to identify the common HR problems and see how effective GPT-3, a third-generation human language prediction model, is in solving these issues.
A major finding from the survey was that a significant number of HR professionals believe their tasks could be automated. However, the complexity of certain workplace issues is challenging for AI to navigate.
HR Professionals Feel Their Jobs Could Be Automated
HR deals with some of the most complex issues in an organization. This department is the primary bridge between the workforce and the organization. Having said that, the study showed that 86% of HR professionals felt that it was likely that their jobs would be replaced by automation. In fact, about 28% also felt that AI could very or extremely likely replace them.
Percentage of HR professionals thinking they could be replaced
Source: Skynova surveyOpens a new window
Several HR-related issues are complex and pose a risk to the organization. For example, employee salaries, health and mental wellness, and poor management are the top three issues for HR professionals. About 23% of respondents said employee salaries posed a significant threat, and the same percentage said employee wellness problems posed a threat to the company. About 30% said the problem with leadership would be a significant threat.
While a vast number of professionals said their jobs would likely be replaced by automation, many also felt that all these issues might not be appropriate for automating.
HR Professionals Show Mixed Reactions for Automating Issues
To understand the effectiveness of GPT-3 in automating issues, Skynova generated various responses for common problems using GPT-3. The respondents were then asked to rate the responses without telling them that the responses were generated by an AI. While a few AI responses received high satisfaction ratings from the respondents, some did not fare so well. Further, it was interesting to see that while some of the scores the AI responses received were the highest an AI could get, the most popular responses were still rated just above average by the respondents.
Some of the responses that scored high were about harassment, bullying, and diversity and inclusion (DEI), the issues most likely to cause legal problems.
Best and worst AI responses
Source: Skynova surveyOpens a new window
Having said that, certain responses received a low average score from the respondents. Some of these were related to employee pay and salaries and company benefits and/or culture. Respondents felt that the post-pandemic business world had exacerbated employee pay and benefits-related issues. Hence, respondents were not satisfied with AI’s responses. They also thought that the issue related to employee vaccination and return-to-work was complex and could not be automated. The Ai response to this issue received the fifth-lowest score.
The responses and scores they received may indicate that there may be some limitations to the AI’s ability to provide satisfactory responses in certain situations.
Should Organizations Pursue HR Automation?
Given that HR professionals were not completely satisfied with the responses, and some responses may also exacerbate certain issues, is it worth automating HR tasks and processes? About 33% of the respondents said that automation could streamline requests and help them prioritize their tasks. About 34% also felt that the technology could provide quick responses to less complex problems but not employee issues.
At the same time, 41% believed that automation would pose legal risks. About 48% also felt that AI would not understand the complexity of an offense to provide an appropriate response, especially in cases of harassment and misconduct at the workplace. In fact, according to another study by Skynova, 55% of small businesses believe that HR should not be automated due to ethical concerns.
Given that AI can automate a few HR tasks, and in fact, make HR processes more efficient, it should undoubtedly be used wherever possible. However, certain tasks and processes require critical thinking, adaptability according to the case, and a human presence and touch. An automated solution in these cases may do more harm than good. As such, these tasks and processes are best left to humans than machines.