A new AI algorithm can detect whether a person is inebriated simply by listening to their voice
Researchers from Australia’s La Trobe University have developed a deep learning algorithm that can instantly determine whether a person has consumed more than the legal limit based on a 12-second voice recording
It’s a well-known fact that alcohol can affect a person’s speech. Slurred words, a change in volume or pitch, and an overall loss of coordination are all telltale signs that someone has had a few too many.
Now, researchers from Australia’s La Trobe University have created a deep learning algorithm that can instantly assess whether a person has over the legal drinking limit based on a 12-second speech clip.
Using a collection of 12,360 audio recordings of intoxicated and sober speakers, the algorithm was constructed and evaluated. The recordings were made in a variety of environments, including bars and nightclubs, phone calls, and podcasts, and included a diverse range of accents.
According to the researchers, one of the potential future applications may be integrated into mobile applications and used in diverse contexts to obtain fast results concerning people’s levels of drunkenness.
This could be useful in a variety of situations, such as law enforcement stopping someone on suspicion of drinking under the influence or employers conducting random breathalyzer tests on staff members.
The algorithm could also have implications for the taxi and hospitality industries. For example, companies could use it to screen patrons for intoxication, and clubs could use it to refuse entry to people who are already intoxicated.
The algorithm is still in the early stages of development, and the researchers say that more work is needed to improve its accuracy. However, they are hopeful that it will one day be able to help people make better decisions about their alcohol consumption.
Read more about AI:
Any data, text, or other content on this page is provided as general market information and not as investment advice. Past performance is not necessarily an indicator of future results.
The Trust Project is a worldwide group of news organizations working to establish transparency standards.