A new study from Stanford University has found that tongue-twisters could be used to gauge alcohol-intoxication levels in people. The researchers say that the changes in voice patterns caused by alcohol consumption could be detected by smartphones or smart speakers, and potentially prevent drunk driving or other risky behaviors.
The study involved 18 adults aged 21 and over, who were given a weight-based dose of alcohol and randomly assigned a series of tongue-twisters. They had to recite one tongue-twister before drinking, and one each hour up to seven hours after drinking. Their speech was recorded by a smartphone placed on a table about two feet away, and their breath alcohol concentration was measured every 30 minutes.
The researchers used an artificial intelligence system to analyze the voice features related to pitch and frequency, and to predict the alcohol-intoxication levels based on the speech data. They found that the voice analysis was 98% accurate in detecting intoxication, defined as a breath alcohol concentration of more than 0.08%, which is the legal limit for driving in the US.
Why tongue-twisters are useful
The researchers chose tongue-twisters as a way to stress the vocal system and bring out the changes that might not be detected when speaking normal sentences. They say that tongue-twisters are challenging for sober people, and even more so for intoxicated people, who may have difficulty with articulation, coordination, and memory.
Some of the tongue-twisters used in the study were:
- How can a clam cram in a clean cream can?
- She sells seashells by the seashore.
- Peter Piper picked a peck of pickled peppers.
The researchers say that tongue-twisters could be used as a form of voice challenge that would not allow someone to start their car or operate heavy machinery unless they pass the test. They also suggest that tongue-twisters could be used to support bartenders or restaurant staff in deciding when to cut someone off from drinking.
What are the potential applications
The researchers say that their study is a proof-of-concept that shows the feasibility of using voice analysis to detect alcohol intoxication. They envision that this technology could be integrated into smartphones or smart speakers, and used to monitor people’s drinking behavior and provide feedback or intervention when needed.
For example, the researchers say that the technology could be used to:
- Send harm prevention messages or reminders to drink water or eat food when someone is drinking.
- Alert a friend or family member when someone is drinking too much or needs help.
- Suggest alternative transportation options or offer to call a taxi when someone is too drunk to drive.
- Provide information or resources on alcohol use disorder or treatment options when someone is drinking excessively or frequently.
The researchers say that their goal is to develop a tool that could help people make informed and safe decisions about their alcohol consumption, and prevent alcohol-related harms such as drunk driving, violence, or injury.
What are the limitations and challenges
The study has some limitations, such as the small sample size, the lack of diversity in the participants, and the controlled laboratory setting. The researchers say that more studies are needed to validate their findings in larger and more diverse populations, and in real-world scenarios.
The researchers also acknowledge that there are some ethical and social challenges in using voice analysis to detect alcohol intoxication. They say that the technology should be used with the consent and awareness of the users, and that the privacy and security of the voice data should be protected. They also say that the technology should be used as a complement, not a substitute, for human judgment and intervention.
The researchers say that they hope that their study will inspire more research and innovation in using voice analysis to detect and prevent alcohol-related harms, and to promote healthy and responsible drinking.