The Uttar Pradesh Police is compiling a database of voice samples of accused in cybercrime, extortion, kidnapping for ransom and other serious offences. The database will help investigators match suspicious audio recordings with stored samples and identify suspects operating through phone calls and digital platforms.
Under the new facility on the police’s Yaksh app, investigators will upload voice samples of accused linked to specific cases. The system aims to streamline voice searches, compare samples with suspicious recordings and use artificial intelligence-based analysis to improve crime detection. The process begins with investigators uploading an accused’s voice sample to the Yaksh app along with the prescribed case details. Cyber police stations will handle recordings linked to cybercrime, while district police stations will upload samples related to cybercrime, extortion, kidnapping for ransom and other serious offences. For cybercrime cases, the recording can be linked to an FIR or a National Cyber Crime Reporting Portal (NCRP) complaint number. Investigators must enter details such as the FIR date, district and police station before uploading the audio. For extortion, kidnapping for ransom and other serious offences, officers can select the relevant crime category and upload the voice sample. An FIR number is not mandatory under this option. The DGP has cautioned officers against uploading a complainant’s voice as that of an accused. Station house officers, cyber police station in-charges and designated district-level supervisory officers will monitor compliance and ensure that recordings meet the prescribed standards.
In a circular issued on September 30, DGP Rajeev Krishna directed cyber police stations and district police stations to prioritise uploading quality voice samples, with supervisory officers responsible for ensuring that the recordings are accurate and correctly attributed to the accused. The circular mandates a clear voice recording of at least 30 seconds, along with the required remarks before the sample is saved. If a recording contains both voices, investigators must extract only the accused’s portion and prepare a clip of at least 30 seconds.
