Scientists from the Stanford Institute for Artificial Intelligence have created an algorithm that lets you predict pay level based on job descriptions. VentureBeat writes about it.

Researchers collected more than a million employee search records to compare the relationship between ad copy and salary. They then used the natural language processing model. BERT and with its help 800,000 vacancies were filled.

In the course of testing the system on the remaining records, it turned out that the algorithm predicted the match between job description and salary with 87% accuracy. By comparison, using only ad headlines and geographic locations only produced accurate predictions 69% of the time.

Using the algorithm, the scientists also identified a relationship between the availability of certain certifications and expectations for salary increases. They believe that having such information will help employees choose truly necessary training courses.

Lead researcher Sarah Bana stated that not only applicants will benefit from the algorithm. Employers can use this information to better invest in human capital and retrain specific employees early, he says.

For authorities, the data will help support vocational training programs and understand the economic value of skills, he added.

“Tools like this will help job seekers improve job prospects, employers improve workforces, and governments respond to sudden changes in the economy,” the researcher said.

In the future, Bana hopes to develop a web-based application where employees or companies can discover added value by upgrading their skills or moving to a new location.

The scientist will also try to characterize the contribution of various words to the salary estimate.

“Ideally, we would color the words from red to green in messages where dark red words are associated with low pay and dark green with high pay,” he says.

Recall that in May, Singaporean authorities talked about the massive promotion of robot assistants by local businesses due to labor shortages.

In June 2021, scientists found that automation and artificial intelligence have led to a 50-70% drop in US wages over the past 40 years.

Follow ForkLog AI on TikTok!