COVID has increased the demand for inmate telephony to an unprecedented degree. One UK prison reports a surge in demand of 600% on pre-COVID levels as visiting restrictions take hold.
Yet the need to perform targeted and random monitoring remains the same. A 6x increase in calls means a 6x increase in prison staff time spent listening to calls. This can mean listening to 3 or 4 hours of audio to find 2 minutes of interest and is paid for by increasing overtime or cutting front-line staffing.
IV has the answer to help better identify calls to monitor, and to help speed up review.
Using its hyper-accelerated speech recognition algorithms, IV can speed through thousands of hours of prisoner telephony and video in no time.
It then uses a blend of Natural Language Technology and sophisticated search to help pinpoint calls that should be listened to, which allows better targeting and quicker review.
The system can easily be tuned to adapt to new slang, words or phrases that are of particular interest. It presents the spelling of the most likely language of the most likely spoken word while retaining a full index in all processed languages for search purposes. Ongoing research work means that more and more dialects and accents can be detected over time, and the system can operate in over 25 languages and dialects, automatically selecting the best one for every call.
Intelligent Voice (IV) has a longstanding record of success in delivering security and intelligence-applicable technologies. IV’s Graphics Processing Unit (GPU) accelerated ASR and SmartTranscript solution, which builds on ASR, provides a faster, better assessment of selected audio.
The solution is created for how people speak rather than how they type. It reviews vast numbers of recordings through pre-highlighting of transcribed text via patented Natural Language Processing (NLP) for speech, as well as listening to a human voice, thereby rapidly assessing mission-intelligence and value in a call.
The audio review player incorporates the audio file, linked or embedded, with a navigable time-bar waveform, playback rate slider, and jump back and forward controls – skip through silence.
Organisations can identify calls of interest, based on both keywords they define and topics the NLP exposes, and prioritise those calls for further review through clickable links within the player that navigate to the related parts of the audio.
Every call into or out of US prisons is recorded. But the recordings generate huge quantities of audio that are prohibitively expensive to monitor with human ears. To help, one jail in the Midwest used a machine-learning system developed by Intelligent Voice (IV), a London U.K. based firm, to listen in on the thousands of hours of recordings generated every month.
The Intelligent Voice software analysed the phrase “three-way” appearing numerously in phone calls – it was one of the most common non-trivial words or phrases used. The software assessed that it was a code. Prisoners are allowed to call only a few previously agreed numbers. So if an inmate wanted to speak to someone on a number not on the list, they would call their friends or parents and ask for a “three-way” with the person they really wanted to talk to i.e. a code for dialling a third person into the call. No one running the phone surveillance at the prison spotted the code until the IV software started churning through the recordings.
Intelligent Voice continues to lead the market and maintain its strengths in the areas of thought leadership, innovation, research and providing solutions. This case study illustrates that commitment and the speed, scale of analysis that IV’s machine-learning algorithms are bringing to the world.