Start practicing your tongue twisters. Artificial intelligence could soon judge your sobriety based on how well you can recite tongue twisters.At least some researchers suggest so after research the study Analyzed intoxication levels based on audio with amazing accuracy.of guardian We will explain in detail the findings of the paper published in . Alcohol and Drug Research Journal Early this month. In what may be remembered as the most fun experiment to take part in, 18 adults of legal drinking age were given vodka gimlets until they became inebriated. Participants were then asked to recite a tongue twister every hour, and their breath alcohol levels were recorded at 30-minute intervals.
The researchers trained an AI to look at changes in voice pitch and frequency at different levels of intoxication and analyze the results. As a result, the program was able to predict with 98% accuracy whether someone was within the legal drink-driving limit. “With the proliferation of smartphone sensors, we are now able to use digital signals to more accurately predict when a drinking episode will occur, enhancing our ability to intervene at the most effective moments,” the study said. said lead author Dr. Brian Suffolette. register. Suffolette, an associate professor of emergency medicine at Stanford University, believes developing some real-world applications is easy enough.
“The most obvious one is some type of ignition lock on your car. [that] It won’t allow the car to start unless it can pass a “voice challenge,” which could be used to ensure public safety in certain high-risk workplaces, such as school bus drivers and heavy equipment operators. “There is,” he said. guardian. He added that restaurants and bars could make available devices to control when patrons are blocked from purchasing additional drinks. Although the technology behind this concept is interesting, this study included only a small sample and racial composition (all participants were white). “We believe there is the potential for exciting developments that could ultimately be really useful,” said Petra Meyer, professor of public health. “But obviously you’ll want to test this approach on a larger and more diverse sample first.” (more AI story).