In an era where technology is ubiquitous, scientists have made a groundbreaking discovery that revolutionizes how we monitor our alcohol consumption.
A research team from Stanford Medicine and the University of Toronto conducted a study showing that changes in voice after drinking alcohol can be a reliable indicator of whether a person is intoxicated.
Smartphones can assess blood alcohol levels while we drink. (Illustrative image: Les Numériques).
This discovery paves the way for the use of smartphones as tools for monitoring alcohol levels.
The study involved 18 adult participants, who were given a quantity of alcohol based on their body weight. After drinking, the participants were recorded reading a tongue twister (a complex sentence) before and after consuming alcohol, using a smartphone positioned 30-60 cm away.
Their blood alcohol levels were also monitored by scientists every 30 minutes throughout the 7-hour study.
The voice recordings were divided into one-second segments and analyzed using metrics such as pitch and frequency. Thanks to advancements in signal processing, audio analysis, and artificial intelligence, this model can accurately predict intoxication levels up to 98%.
Brian Suffoletto, an assistant professor of emergency medicine at Stanford University, expressed surprise at the model’s accuracy.
“The results truly surprised us. Although this study is not pioneering in highlighting changes in speech characteristics when someone is drunk, we are confident that its superior accuracy stems from applying advanced techniques in signal processing, acoustic analysis, and AI,” said Suffoletto.
Implications and Future Applications
This technology could be integrated into smartphones, allowing users to monitor their level of intoxication and send alerts when someone is too drunk to drive.
Suffoletto also suggested integrating this system with other phone features, such as using accelerometers to check for unusual walking patterns or text message analysis systems to detect changes in communication styles that may indicate drunkenness.
The research underscores the importance of technology in preventing dangerous behaviors related to alcohol. By providing real-time intervention capabilities, these tools have the potential to save lives by preventing intoxicated individuals from getting behind the wheel.
The researchers plan to expand their database with more voice samples from additional participants to further improve their model.
