Pindrop’s Pulse Inspect detects AI-generated speech in audio files with 99% accuracy, helping media outlets combat misinformation and protect their credibility.
Pindrop, a global leader in voice authentication and security since 2011, today announced the launch of Pindrop® Pulse™ Inspect in Preview. This innovative tool, the latest addition to Pindrop’s cutting-edge deepfake detection Pindrop® Pulse™ product family, can detect AI-generated speech in any digital audio file with 99% accuracy1. Following the launch of Pindrop Pulse earlier this year and Pindrop’s latest financing in July, Pindrop Pulse Inspect extends Pindrop’s technology beyond just call centers but now offers a new use case for media organizations, nonprofits, government agencies, and social media networks. Pulse Inspect quickly analyzes and verifies whether media files contain synthetic speech so customers can proactively prevent the spread of misinformation and maintain their brand credibility.
Generative AI has led to a rise in deepfakes, posing a serious threat to organizational credibility. Pindrop’s Deepfake and Voice Clone Consumer Sentiment Report estimates that over 12 million U.S. adults know someone who has personally had deepfakes created without their consent. These synthetic files are often spread through platforms like YouTube, TikTok, Instagram, Facebook, and news media. Organizations can adopt this cutting-edge technology to help prevent exposure to these evolving threats to safeguard their public image and credibility. Pindrop’s Pulse solution tackles this issue head-on by accurately detecting deepfakes.
Also Read: Why Marketers Should Lead the Generative AI Revolution
By adopting Pulse Inspect, users can upload audio files into the Pindrop platform to quickly and accurately determine if the files contain synthetic speech. Pindrop provides immediate deepfake scores and visibility into specific parts of the file that are identified as containing synthetic or deepfake artifacts so the user can pay closer attention to that segment when reviewing and conducting the analysis. With Pulse Inspect, users can take swift action to safeguard the integrity of content before distributing it on any online public news, social, or publishing platform.
“For over a decade, Pindrop has been at the forefront of voice authentication and security innovation. For example, we discovered the source TTS engine used in the Biden deepfake incident earlier this year, and the critical need for robust deepfake detection mechanisms has come through loud and clear in my conversations with dozens of stakeholders across industries,” said Pindrop co-founder and CEO, Vijay Balasubramaniyan. “Today, we are launching Pulse Inspect to extend our expertise in deepfake detection to organizations like media, nonprofits, governments, and social media networks at a pivotal time in history. With Pulse Inspect, we’re empowering organizations to safeguard the integrity of their content and bring trust back to the modern digital age.”
Pulse Inspect is powered by Pindrop’s advanced deepfake detection technology, which NPR reported as 40 percentage points more effective than the nearest competitor in an independent study. Pindrop has trained its deepfake detection model on 350+ deepfake generation tools, 20M+ unique utterances, and 40+ languages, covering over 90% of the internet’s spoken languages3. Continuous deepfake scores are provided every four seconds, allowing the identification of synthetic versus human content in mixed media. Tracking the liveness continuously enables Pulse Inspect to detect which audio segments are synthetic versus genuine, which is particularly helpful when only part of the audio is a deepfake. Its easy-to-use, API-driven batch-processing platform and simple user interface provide clear analysis (synthetic vs. non-synthetic) with adjustable thresholds based on risk tolerance, ultimately saving time, money, and resources.
Also Read: AI’s Content Conundrum: Fake, Fast, and Dangerous
“As a leader in synthetic voice for Hollywood, Respeecher established guidelines and policies that ensure our technology is not misused, and we remain committed to pioneering ethical development and use of AI voice cloning technology,” said Alex Serdiuk, CEO and Co-Founder of Respeecher, a partner of Pindrop. “Pindrop’s Pulse Inspect technology helps us ensure that our detection technology stays ahead of generative AI attack threats. We are excited to see the launch of Pindrop Pulse Inspect and look forward to putting the technology in the hands of our joint customers so they can forensically examine any audio and determine its synthetic provenance.”
“With the upcoming US elections and dramatic growth in synthetic AI-generated audio appearing in these calls in 2024, YouMail, a visual voicemail and Robocall blocking service, recognizes Pindrop Pulse Inspect as the go-to solution to accurately detect deepfakes in voice calls,” said Mike Rudolph, Chief Technology Officer at YouMail. “Pindrop’s technology was resilient to different audio formats and impressively accurate at scale. We are excited to partner with Pindrop to identify the election-related robocalls that are deepfakes, especially as the volume of these calls ramps during the critical US election season.”
“TrueMedia.org is a non-profit, non-partisan AI project to fight disinformation in political campaigns by identifying manipulated media,” said Oren Etzioni, CEO of TrueMedia.org. “Our comprehensive evaluation found Pindrop’s audio deepfake detection has better accuracy than other alternatives in detecting synthetic speech. We are excited to partner with Pindrop in this mission and add Pindrop’s deepfake detection technology to the solution for our customers and users worldwide.”
