Thapar students win 2 lakh rupees for a deepfake detector: defense is also taught in the classroom

By Urban Acres · October 25, 2026.

According to a headline in Urban Acres, a team of Thapar students has won a ₹2 lakh (200,000 Indian rupees) prize for an artificial intelligence system that detects deepfakes. A deepfake is a video, audio clip or image that has been generated or manipulated with AI to look real.

All we have is the headline and a short description. We don't know what technique the system uses, what data it was trained on, how accurate it is or which competition it won. So we won't describe how it works or judge its quality, because that would mean making things up.

Our take is more modest, but still useful. University prizes rarely lead to products that reach the market, and deepfake detectors have a well-known weakness. A detector trained on today's generation techniques usually becomes less effective as new generators appear. Creators and detectors are locked in a constant race, and lab results rarely hold up in the real world without regular updates.

Even so, the story matters for what it signals. Fraud using synthetic voices and faces is one of AI's most concrete near-term risks, and it's a good sign that students are already working on defenses. It would be worse for society if only big companies had that capability.

In the long run, the answer will probably go beyond spotting fakes to certifying what's authentic, with cryptographic signatures of origin built into cameras and platforms. Detectors will be one layer of that system, not the whole solution. If the team publishes its results and methods, we'll revisit the story with data.