Pindrop Pulse offers protection against audio deepfakes
Pindrop launched Pindrop Pulse, an audio liveness detection capabilities for real-time identification, monitoring, and analysis of audio deepfakes.
Notably, Pulse was instrumental in identifying the TTS engine used in the recent President Biden robocall attack, By leveraging advanced deep learning models for predictive assessments between real and synthetic voices, this solution is engineered to address existing fraud threats and proactively prepare for the dynamic evolution of fraudulent activities.
As we stand on the brink of what is projected to be a seismic shift in the landscape of cyberattacks—with potential damages expected to soar to $10.5 trillion by 2025—the launch of Pulse emerges as an essential, perfectly timed defense in this critical moment. Amidst a climate where the sophistication of cyber threats is advancing at an alarming rate, this innovation addresses the paramount importance of forward-thinking technological defenses.
The efficacy and reliability of Pulse have been endorsed by leading financial institutions, including First National Bank of Omaha (FNBO), which has been utilizing Pulse in beta to bolster its customer service security measures.
Steve Furlong, Director of Fraud Management at FNBO, stated, “In an era where AI advancements bring both innovation and new threats, FNBO remains committed to protecting our customers and their information. In an effort to proactively combat the emerging threat of deepfakes, our partnership with Pindrop provides us with cutting-edge solutions that safeguard our customers’ information with precision. After rigorous testing, we’re very happy with the results – Pindrop’s technology ensures our defense mechanism is robust against advanced threats. Their commitment to excellence and innovation makes them an invaluable ally in our mission to protect our customers.”
In Pulse’s beta testing phase, it demonstrated remarkable efficacy in voice security, pinpointing manipulated voices with just two seconds of audio and offering responses in a swift 150 milliseconds. This rapid detection and response capability allows call centers to swiftly neutralize threats in real time.
Engineered to counteract four predominant types of audio fraud—replayed voices, speech synthesis, automated voice chatbots, and real-time voice conversion—adeptly mitigating the risk of audio deepfakes. Beyond its detection capabilities, Pulse enhances Pindrop’s broader suite of solutions, offering seamless integration with existing authentication and fraud prevention services to fortify a holistic defense against voice communication threats.
“Pindrop Pulse represents the company’s evolution in response to increasing sophistication of attackers and allows us to answer the fundamental question of whether we are interacting with a real human or a machine” said Vijay Balasubramaniyan, CEO of Pindrop. “Our solution doesn’t just detect deepfakes; it is a pivotal step towards reinstating trust in commerce, media, communication, and society as a whole. We’re excited to be at the forefront of this charge, equipping our clients with the necessary defenses against the sophisticated cyber threats that Generative AI has created.”