Can AI Measure Human Skills?

A well-known driver of education is to help people develop skills and drive human capital. The second element of education is its role in signalling skills to the labor market. The successful bridge between education and work relies on companies understanding what skills matter to their productivity and what skills are achieved within education.

With the increasing pace of technology and lifelong learning, educational institutions are moving towards becoming insufficient in their ability to selectively signal skills. The scaling of online education also allows mass accessibility to high quality education in a way that traditional brick and mortar education can not. However, without selectivity, the question remains as to whether it can compete with traditional education credentials and it’s signalling value. Answering this will rely on whether we can credibly measure the skills that people have and how that translates to productivity in the role.

The hiring process is often reliant on a single piece of paper however, increasing forward thinking companies are realising the long-term challenges in relying on one source of information and so have moved to tech-enabled solutions to help create a more holistic profile of candidates that puts data first.

One example is DBS Bank. The talent acquisition team created JIM (Jobs Intelligence Maestro) that could conduct candidate screening for wealth planning managers, a high volume job in the consumer bank. JIM was able to shorten candidate screening from 32 minutes to 8 minutes, improve the job application completion rate to 97% and respond to 96% of candidate queries. Hilton is another company which accelerated their speed to hire by 85% through using AI in sourcing, screening and interviewing candidates in their hiring process.

Adapting to skills-based hiring will lead to altering job descriptions and potentially the removal of a degree requirement altogether. However, this will also require a mindset shift on how companies select skills. Across all industries, adopting a skill-based hiring approach will need to identify benefits and barriers of widening the talent pool and redeveloping strategies in the hiring process. Focus will also lead to educational and learning pathways to upskill a more diverse employee population.

The challenge in using AI is to successfully source and measure skills. It can be argued that measuring skills is not required to measure a skill, rather measure a task. Because of inability to measure a task at a high level we find proxies to help, these being the skills. For example, in a football game players can be judged on their ability to score a goal. However, measuring their performance can only be achieved in matches. These are often scarce and players are generally on the bench rather than on the pitch. Therefore, how can you continue to assess the player’s abilities when they are not giving the opportunity to play? Rather, their skills are measured in terms of shooting with accuracy and speed, assisting goals and fitness. Our intuition tells us that their ability to perform these different unique skills correlates to the striker being able to score goals in a match. These individual skills are easier for a human and also a computer to measure and assess.

There is still a long way to go for computers to successfully build this type of assessment on individual skills and link them to tasks. The ability to measure soft skills over hard skills and taking hard skills apart to determine a common language in assessing candidates needs to be achieved first before they can be accurately measured. However, it seems co-bots continue to infiltrate the future of work…