Artificial intelligence is becoming increasingly adept at identifying illnesses based on voice patterns, as demonstrated by new research from the Sardar Vallabhbhai National Institute of Technology in Surat. The study found that the tone of a person’s voice can be analyzed by AI to identify a cold, making it more difficult for people to falsely call in sick when they are actually fine. The research team examined the voice patterns of 630 people, 111 of whom had a cold, to determine who had the illness.

The Economist reported that the study utilized voice patterns that were observed in human speech, which can be thought of as harmonics that decrease in loudness as their frequency rises. When a person has a cold, they may also exhibit an irregular pattern of the same harmonics. Based on this phenomenon, the researchers used machine-learning algorithms to analyze the amplification differences of various individuals and identify those who had a cold.

In the study, participants were asked to count from 1 to 40 before reporting what they did over the weekend, then reciting Aesop’s tale The North Wind and the Sun. The researchers were able to identify colds with 70% accuracy.

While the study was intended to determine whether a person had a cold without needing to see a doctor, it may also be of interest to businesses that want to catch employees who falsely claim to be sick to take time off from work. The study’s results suggest that AI could be a useful tool in this regard.

However, the increasing use of AI in this way raises concerns about privacy and employment. It remains to be seen how companies will implement AI technology to monitor the health of their employees, and how employees will react to being scrutinized in this way.