New Deep Learning AI Can Detect Alcohol From Your Voice
Researchers at LA Trobe University have created an AI-based alternative for Breathalyzer tests.
The newly developed technology has been dubbed the Audio-based Deep Learning Algorithm to Identify Alcohol Inebriation (ADLAIA). According to a paper published in the journal Alcohol, the algorithm can determine an individual’s intoxication level based only on a 12-second recording of their voice.
The algorithm was developed using a dataset of 12,360 audio clips from inebriated and sober speakers. According to the paper, ADLAIA could identify those with a blood alcohol content (BAC) of 0.05% or higher with almost 70% accuracy, and those with a BAC of 0.12% or higher with almost 76% accuracy.
For comparison, Breathalyzers have a margin of error between 15% to 20%.
“Being able to identify intoxicated individuals solely based on their speech would be a much cheaper alternative to current systems where breath-based alcohol testing in these places is expensive and often unreliable,” remarked researcher Abraham Albert Bonela.
Given the inaccessibility of BAC tests to everyday consumers (the average Breathalyzer runs the gamut from $40 to $130), ADLAIA could level the playing field and provide an all-around low-budget alternative.
“Upon further improvement in its overall performance, ADLAIA could be integrated into mobile applications and used as a preliminary tool for identifying alcohol-inebriated individuals.”
Soon enough, you may be speaking into a police officer’s phone next time you get pulled over on the road.
This post may contain affiliate links, so we may earn a small commission when you make a purchase through links on our site. This helps support Tequila Raiders at no additional cost to you.