As concerns about the impact of artificial intelligence (AI) on work continue to rise, a recent study conducted by researchers at the Sardar Vallabhbhai National Institute of Technology in Surat, India has shown that AI may have the potential to help prevent employees from falsely claiming sick leave when they are actually not ill.
The study, reported by The Economist, focused on analyzing vocal patterns in human speech to identify individuals who were genuinely suffering from a cold.
The research team, led by electronic engineer Pankaj Warule, utilized the vocal rhythms detected in speech, which exhibit harmonics that decrease in amplitude as the frequency increases.
The team hypothesized that the presence of a cold could disrupt this regular pattern. By incorporating these vocal components into machine-learning algorithms, the researchers examined the differences in amplification to determine if a person had a cold or not.
During the study, participants were asked to perform tasks such as counting from one to 40 and describing their weekend activities before reciting the Aesop’s fable “The North Wind and the Sun,” which is commonly used in phonetics research.
Can AI Detect Cold?
The AI algorithm was able to accurately detect the presence of a cold in the participants about 70% of the time.
The researchers primarily aimed to develop a non-invasive and cost-effective method to identify colds without the need for costly and time-consuming visits to the doctor.
However, the findings may also be of interest to employers who are keen to identify employees who may be pretending to be sick to take time off from work.
This could potentially involve asking employees to read a fable as part of the detection process.
While there is limited research on the prevalence of employees faking sickness to avoid work, it is known that in the United States, many private sector employees do not have paid sick leave.
In fact, nearly a quarter of private sector employees in the US lack access to paid sick leave, unlike in many other countries where paid sick leave is mandated by the government.
This development in speech recognition using AI is part of the ongoing trend of AI being applied in the medical field for various purposes. For instance, in a recent study, AI was used to detect Parkinson’s disease through analyzing breathing patterns of patients.
Additionally, AI-trained analysis of vocal patterns has shown potential in detecting depression and several forms of cancer.
In conclusion, the use of AI to analyze vocal patterns in speech has shown promise in identifying the presence of a cold and may have potential applications in preventing employees from falsely claiming sick leave.
However, further research and validation are needed before widespread adoption of this technology.