Sensors in smart phones and smart speakers may soon be used to determine a person’s level of alcohol intoxication based on changes in their voice, as per a study published in the Journal of Studies on Alcohol and Drugs.
THE STUDY
Researchers from Stanford Medicine and The University Of Toronto conducted a small-scale study involving 18 adults aged 21 and above. Participants were administered a weight-based dose of alcohol and asked to read aloud a series of tongue twisters—one before drinking and one each hour for up to seven hours after consumption. While recording their voices, the researchers also measured the participants’ breathe alcohol concentration at regular intervals.
By analyzing measures like frequency and pitch in the recorded voices, researchers were able to predict alcohol intoxication with 98% accuracy when compared with breath alcohol results.
THE IMPLICATIONS: JUST-IN-TIME INTERVENTIONS
Dr. Brian Suffoletto, the lead researcher and an associate professor of emergency medicine at Stanford, expressed surprise at the model’s accuracy and stated that the goal is to provide “just-in-time interventions” to prevent accidents caused by intoxication. Given the widespread use of smart phones and smart speakers, they are ideal tools for alerting people when they become intoxicated.
FUTURE PREDICTIONS: INTEGRATED SURVEILLANCE TOOLS
Dr. Suffoletto envisions future surveillance tools integrating multiple sensors—analyzing gait, voice, and texting behaviour— to provide a more comprehensive and continuous data input. However, he insists on the need for larger studies involving a diverse range of ethnic backgrounds to validate voice patterns as a reliable indicator of intoxication. He also recommends collaborations with companies already collecting speech samples through smart speakers.
The ultimate goal is to develop an intervention system that people are willing to use, which can help prevent injuries and save lives. As Dr. Suffoletto notes, the timing of interventions is crucial. While reminders of consumption limits can be impactful at the start of a drinking episode, their efficacy decreases as the person becomes significantly intoxicated.
