SRI’s Speech Technology and Research Lab began developing new AI-driven approaches to real-world, operational speech processing under DARPA’s Robust Automatic Transcription of Speech (RATS) program.
While commercial software solutions focus on clean recordings and highly controlled environments, speech processing in defense, intelligence, and law enforcement often requires accommodating degraded speech signals from highly variable and uncontrolled environments.
To fill this gap, SRI further matured technologies developed via the RATS program and transitioned them through the Open Language Interface for Voice Exploitation (OLIVE) platform, an SRI-operated platform that now provides customers in the defense community, the intelligence community, and federal law enforcement with numerous mission-critical speech processing capabilities.
The OLIVE platform offers both real-time and post-mission speech processing for multiple mission types. The platform operates across languages of high operational interest with capabilities including robust dialect identification, voice biometrics and speaker tracking, and language-independent keyword detection. Leveraging technology developed through subsequent DARPA programs like Semantic Forensics and Media Forensics, OLIVE also provides image/video analysis, face biometrics, and synthetic/manipulated media detection capabilities critical to federal law enforcement activities.
Beyond supporting government customers, OLIVE has delivered cutting-edge speech processing capabilities to multiple commercial startups, including Pienso (an interactive AI-driven data platform) and Kasisto (agentic AI for banking).
