AI & ATS: The Good, The Bad & The Ugly

From gaming and self-driving cars to Siri and Alexa, artificial intelligence (AI) impacts almost every aspect of our daily lives. Researchers continue to push boundaries on how to apply this technology. It’s no surprise to find AI used more frequently to streamline the recruitment and hiring process. While this provides benefits, it also comes with some major flaws.

The Good

Applicant tracking systems (ATS) have been used for years to help résumé-swamped recruiters and hiring managers quickly sort through hundreds – sometimes thousands – of applications for a single job opening. Acting as the first line of screening, ATS allows companies to develop short-lists of qualified candidates.

With the implementation of advanced AI solutions, ATS has received a big boost in reliability. Also, it has continued to help automate time-consuming HR processes. This is saving not only time in making hiring decisions, but also money by potentially avoiding bad hires. Last year, Google introduced its own AI-driven ATS — Google Hire. It uses information from past hires to improve ATS functionality. So, what are the downfalls of AI and ATS?

The Bad

Generally speaking, ATS uses a subfield of AI known as machine learning. The system applies large amounts of data in order to “learn” patterns or categories. In addition, the ATS parses data and uses the learnings to identify patterns of interest. The problem is the ATS’s outcome is only as robust as the data with which it has been given.

As such, Google Hire’s outputs are only as good as its inputs. The most effective candidate selection requires incredibly large amounts of data. Most organizations just don’t have this information. Additionally, even if an ATS is given vast amounts of high-quality data, there may be underlying issues within the data that distort performance.

The Ugly

An excellent example of how AI can go wrong is one company’s decision to abandon a multi-year project to create an AI-powered recruiting engine. It ultimately demonstrated a bias against women. The tech giant provided its AI with 10 years’ worth of job applications. With it being a male-dominated industry, the vast majority of these applicants were male candidates. Inherently, this led the engine to rank applications with words, phrases, and affiliations typically associated with women as less desirable, as they occurred with less frequency.

However, it’s safe to say that the human touch will never completely be eliminated from the recruiting and hiring process. Key indicators that need assessed, such as cultural fit and interpersonal skills of a candidate, are still best left up to humans. What AI does offer is the opportunity to complement the work of recruiters and hiring personnel, enabling them to focus on areas that data cannot accurately capture.

Want to read more about HR trends? Read our CEO’s recent post on AI and digital systems in the workforce.

What Is Artificial Intelligence? Examples and News in 2019, TheStreet
How is AI Revolutionizing Applicant Tracking and Recruitment Process?, HR Technologist
The Digital Transformation Journey of HR: Expert Perspectives, HR Technologist
Amazon Scraps Secret AI Recruiting Tool that Showed Bias Against Women, Reuters