Artificial Intelligence’s superpowers are constantly expanding as researchers develop new applications for machine learning and data analysis. One such application comes from the researchers at the Sardar Vallabhbhai National Institute of Technology in Surat, who have successfully used AI to analyse voice patterns and identify individuals suffering from a cold.
According to a report from The Economist, the researchers at the institute analysed the voice patterns of 630 persons in which 111 of these people were having a cold. The speech patterns were analysed in an attempt to identify persons who were suffering from cold in real.
ALSO READ l YouTube gets offline viewing, four other crucial updates but only these users can access them
The study utilised the concept of harmonics, which are the vocal rhythms in human speech that decrease in amplitude as their frequency increases. Individuals who are suffering from a cold tend to display an irregular pattern of harmonics. The researchers used machine-learning algorithms to analyse these differences in amplification and accurately identify those who had a cold.
Researchers asked the test subjects to perform various vocal tasks such as counting and reciting a fable. They then reported an accuracy rate of 70% in detecting colds using this method.
ALSO READ l China to mandate security reviews for ChatGPT-like AI chatbot services
While the primary goal of the research was to help diagnose colds without seeing a doctor but there could also be another potential application of the technology wherein the employers could use this technology to identify employees who may be falsely claiming to be sick in order to take time off work. This new approach may provide employers with a more accurate way to determine if an employee is genuinely unwell, potentially reducing the number of fake sick days taken by employees.
