Prepare to test your enunciation skills, as researchers are investigating the potential of using tongue twisters to determine intoxication levels. According to a recent study, speech analysis showed remarkable accuracy in detecting intoxication based on pronunciation. The Guardian details the findings published in the Journal of Studies on Alcohol and Drugs.
This involved an entertaining experiment where 18 legally drinking adults consumed vodka gimlets to reach varying levels of intoxication, during which they recited tongue twisters at hourly intervals, with breath alcohol levels recorded every thirty minutes.
By observing changes in voice pitch and frequency as intoxication levels changed, researchers trained artificial intelligence to analyze the data. The AI was successful in predicting legal sobriety limits for driving with an impressive 98% accuracy. Dr. Brian Suffoletto, the study’s lead author, discussed the potential applications for this technology, suggesting that it could be used for ignition locks in cars and in high-risk workplaces to ensure public safety. Suffoletto also mentioned the possible use of this technology in managing alcohol sales at restaurants and bars. While the concept is intriguing, the study’s limitations include a small and racially homogeneous sample size. “There is potential for exciting developments that could be really useful, but one would first want to test this approach in larger and more diverse samples,” commented Petra Meier, a professor of public health. (More AI stories).
