AI app to detect drunk drivers through Cantonese speech analysis in the works

Hong Kong researchers say they are developing what would be the world’s first AI app that detects intoxication through real-time Cantonese voice analysis, helping users decide whether they are fit to drive after drinking.

A team from the Hong Kong Metropolitan University (HKMU) said on Sunday that the self-assessment tool could not only help safeguard road safety but also potentially monitor language abnormalities linked to medication or fatigue.

Funded by a HK$3.4 million (US$434,251) grant from the Smart Traffic Fund, the researchers will collect more than 1,000 minutes of Cantonese speech samples from around 100 participants aged 18 to 60 to establish a post-drinking speech database.

The database will train a multimodal artificial intelligence (AI) language model that comprehensively analyses speech rate, intonation stability, speech pauses, and deviations in grammatical structure.

This will allow comparison of participants’ performance before and after alcohol consumption to assess how drinking affects their cognitive and linguistic abilities.

Participants must pass a health assessment by a doctor, and a nurse will monitor their alcohol consumption.