Picture this: You’re scrolling through social media when a one-minute video featuring Bengaluru City Police Commissioner B Dayananda pops up offering safety tips for Deepavali celebrations.

‘Hello Bengaluru, this is your police commissioner…’ the video begins. But the person on screen isn’t Commissioner Dayananda; it’s his AI (Artificial Intelligence) avatar.

Joint Commissioner of Police (Bengaluru Traffic) MN Anucheth also recently appeared in a similar AI-generated video last month, talking about traffic diversions during a cricket match at M Chinnaswamy Stadium in Bengaluru. The entire reel, including visuals and voice, was created through an AI tool, with the person on screen being Anucheth’s digital avatar.

“We are now doing this on an experimental basis, assessing the pros and cons of the ethical and legal implications. Based on this (the findings), it will be rolled out in a full-fledged manner,” Dayananda told Moneycontrol.

This is probably a rare instance that a police department has used an AI-generated avatar—a digital likeness designed to represent officials on video. With this technology, videos can be created from simple text prompts, featuring avatars that look and sound like real police officials, complete with realistic speech patterns and lip synchronisation.

This rollout underscores how the police authorities in India’s tech capital, Bengaluru, are exploring the use of generative AI in their day-to-day work.

“In the corporate world, AI avatars are even attending Zoom meetings—we’ve reached that stage. In the police force, we may be among the first in the world to experiment with this” he said.

The Guidelinesby1.Purchase only green crackers and celebrate an eco-friendly Deepavali! When buying these crackers, scan the QR code to verify their authenticity before purchasing.2.Avoid lighting fireworks in crowded places or near buildings to prevent accidents, usepic.twitter.com/CHwgcIj61Y
BengaluruCityPolice (@BlrCityPolice)October 31, 2024

The police authorities have partnered with Aeos, a Bengaluru-based startup that operates at the intersection of generative AI and content, creating AI solutions and experiences for businesses.

How does this work?

The process is simple: Police officials enter their message as a text prompt, and the tool generates a video with the avatar delivering it. Their voice samples are captured beforehand, and the police’s social media team, along with the Aeos team, posts the video on social media.

The police commissioner said the service is particularly useful when they need to quickly reach the public about important events, activities, or urgent situations.

“For law-and-order situations, public concerns, or emergencies like terror threats or violence, this allows us to communicate efficiently with the masses,” he said.

“We want to move beyond traditional press releases and make our communication more engaging,” Anucheth told Moneycontrol. “Social media plays a crucial role in reaching the public, and these short information reels – typically around 30 seconds or one minute – capture more attention.”

He explained that, since they are often in the field or outside the office, this tool makes citizen outreach more convenient.

“If you watch the entire video, you’ll see it’s entirely AI-generated. It’s not just my avatar or the voice—it’s everything, down to the graphics. For example, when the Chinnaswamy Stadium is shown with instructions on parking availability, that, too, is AI-generated,” Anucheth said.

He mentioned that this highlights how AI can transform what would otherwise be a plain press release into an engaging visual presentation, perfectly suited for today’s social media audience – particularly those aged 16 to 25, who are more vulnerable to accidents.

He added, “There’s no need for cameras, lighting, a videographer, an editor, special equipment, or a technical team with specialised skills. All we need is the text of what we want to convey, and even a police constable can be trained to produce such videos,”

Concern over deepfakes

To be sure, the use of this technology raises concerns about potential misuse, particularly at a time when deepfakes have emerged as a major threat, both in India and worldwide, with the ability to spread misinformation and manipulate public opinion among other risks.

Anucheth said that safety measures are in place to prevent misuse, given concerns around deepfakes and the use of synthetic media. “The system requires authorisation, and usage is controlled to prevent any unauthorised access,” he said.

Dayananda also acknowledged the ongoing debates surrounding AI’s ethical implications, adding, “As the police force in India’s tech capital, we need to take the lead on this.”

The Bengaluru police have previously used AI across a range of its traffic management and enforcement efforts. This includes streamlining traffic flow and easing congestion with AI-powered signals to reduce delays, monitoring traffic jams, and detecting traffic violations using AI-based cameras.

Known for his tech-savvy initiatives, Dayananda had earlier introduced a fully paperless system (e-office) across all stations and offices, in alignment with the Mission Mode Project (MMP) under the National E-Governance Plan. He launched the first police blog in 2005 during his tenure as the superintendent of police for the Dakshina Kannada district.