Karnataka forensic department turns to digital tools to identify deepfakes, analyse smartwatch data

Deepfake video analysis

Experts say the rise of deepfake technology has made it increasingly difficult to distinguish between real and fabricated video content. Deepfakes, powered by artificial intelligence (AI), can alter a person’s appearance, voice, or actions, creating realistic but fake videos, images, or audio. Various celebrities, politicians, and public figures have already complained about the circulation of such manipulated content on social media. It is also being used for cyber frauds and digital arrests.

Also, read: AI Cops! Bengaluru police pilots AI-generated avatars for social media outreach

DFSL has recently invited a Request for Proposal (RFP) for the supply and installation of a deepfake video analysis tool at the State Forensic Science Laboratory in Bengaluru.  This will help to analyse digital video files, uncover hidden metadata, and trace the device that recorded the video.

“The tool will detect tampering by evaluating the video’s internal structure rather than its visual content, which is critical for identifying manipulated material. It will also trace the editing history, including any software used, and the methods involved in the transmission and alteration of the video,” said a DFSL official.

The tool will also help quickly identify and process child sexual abuse material (CSAM) and offer the ability to triage videos from mobile devices. Its advanced features include categorising the video file based on the originating device and processing software. A reference library of known devices and platforms will ensure that the analysis remains accurate. The tool will work across a wide range of devices, including smartphones, camcorders, CCTV systems, and social media platforms and provides a comprehensive method to authenticate digital videos.

Also, read: Indian govt's C-DAC working on an app to detect deepfakes

Smartwatch data analysis

DFSL has also invited a Request for Proposal (RFP) for the supply and installation of a smartwatch data acquisition and analysis tool at the State Forensic Science Laboratory in Bengaluru, aimed at enhancing its capabilities to analyse data from smartwatches and connected devices, including those from Apple, Samsung, and Garmin.

“These devices store a wealth of personal data, including GPS locations, activity levels, and health metrics, making them valuable in forensic investigations,” said an official.
This tool will support a range of functions, such as bypassing security measures, decoding application data, and recovering deleted information from smartwatches. It will also allow for data extraction from other connected devices like mobile phones, tablets, and GPS devices he said.

The tool will also support the analysis of various types of data, including contacts, messages, GPS locations, and multimedia files. It will also categorise images into different categories, such as drugs, weapons, documents, money and nudity, using machine learning.

“One of the major advantages of this tool is its ability to connect and analyse multiple devices simultaneously, which will speed up the data retrieval process. It will also be compatible with both networked systems and standalone devices, ensuring flexibility in its use for forensic investigators,” the officials said.

Also, read: Karnataka online trading frauds: Losses skyrocket from Rs 23 crore in 2022 to Rs 903 crore by mid-2024

Drone forensic analysis tool

As drone usage increases across various sectors, DFSL will also use a specialised forensic tool to investigate drone-related crimes. This tool will be used to extract and analyse data from drones, including flight logs, GPS data, and media content stored in drone SD cards or USB drives.

“The drone forensic tool will be capable of extracting and analysing data from both iOS and Android devices and will allow for flight log analysis specifically for Unmanned Aerial Vehicles. It will also support the recovery of deleted files and enable simultaneous data acquisition from multiple storage devices, such as SIM cards, SD cards, and USB thumb drives,” said an official.

The tool will also offer capabilities for imaging, cloning, wiping, and formatting flash media. “This will help us to access key evidence from drones, which have become important in various crimes, from surveillance to smuggling and illegal surveillance,” said an official.

Mobile acquisition and analysis tool

DFSL is also rolling out a mobile acquisition and analysis tool designed to extract forensic evidence from smartphones, tablets, GPS devices, and drones. This tool will support both iOS and Android devices, covering over 32,000 device profiles and 12,700 app versions. “It will also allow us to detect devices automatically or manually by searching via manufacturer, model, IMEI, or other identifiers. The tool’s features include the ability to bypass security locks, extract data from various apps, and even recover deleted information. It will also support advanced extraction techniques such as bypassing encrypted bootloaders and decoding data from popular apps like WhatsApp, Signal, and Facebook,” said an official.

The mobile tool will offer comprehensive analysis capabilities, allowing for tracking of location data, monitoring cryptocurrency transactions, and handling cloud data extractions from over 60 platforms. It will also support more advanced tasks like extracting data from IoT devices and creating secure backups for sensitive information.

“This mobile tool is designed to handle a variety of data types, from contacts and call logs to multimedia files and social media interactions. Its ability to categorise data using machine learning algorithms will improve the speed and accuracy of investigations, especially when dealing with large volumes of data”, the official said.

Also, read: In a first, Karnataka police to deploy AI tools for fund trail analysis, dark web monitoring