Last month, when US President Donald Trump was taking photos at the Taj Mahal in India, the country’s basic was going through a agitated time with common riots. As the police grappled to ascendancy the situation, 53 people died and many more were injured.

Yesterday, India‘s Home Minister, Amit Shah, said during the analysis afterward the riots, the government had identified 1,100 rioters using facial acceptance systems.

More importantly, he noted cops used data from voter ID and active authorization databases to identify insurgents:

We are using face-recognition software to analyze people behind the violence. The software has articular 1100 people, 300 of these people came from UP (Uttar Pradesh state) to carry out violence.

According to a report from The Hindu, a senior police administrator said authorities also used India’s civic agent database, e-Vahan, to catch perpetrators:

Our cadre were also recording videos when the riots were on. Once we had clear video footage, the software was used to match the photographs with that in the database of EC and e-Vahan to analyze the suspects. Basically, even if addition changes his appearance, the chin and the eye do not change. These are the two key facial appearance that will help us nab the suspects

However, Delhi police ahead said they only akin faces with the bent and missing bodies databases. They also claimed not to be using the system to analyze protestors.

If the Indian government uses assorted non-criminal databases, there are apropos the system might be used to snoop on citizens and people assuming bone adjoin authorities.

Read next: 10 tips to help your business acquaint throughout the coronavirus crisis