For the second 12 months in a row, an AI-based safety startup took the prize for Most Progressive Startup at RSA Convention’s Innovation Sandbox competitors.
Final 12 months, HiddenLayer began its presentation with a deepfake video of its CEO asking the judges to not decide it because the winner. This 12 months, Actuality Defender confirmed off its software for figuring out deepfakes and different synthetic content material to assist restore confidence in what we see and listen to.
Deepfakes by way of synthetic intelligence (AI) could also be “one of the consequential issues we face, particularly in an election 12 months,” stated Niloofar Razi Howe, working companion at Capitol Meridian Companions and one of many judges for the Innovation Sandbox competitors. The significance of that mission could have helped put Actuality Defender excessive.
How Actuality Defender Defends Actuality
Actuality Defender’s deepfake detection platform and API lets groups establish fraud, disinformation, and dangerous content material in actual time — utilizing AI to defeat AI, stated Ben Colman, the corporate’s co-founder and CEO.
“Whereas the outputs are pretend, the threats are actual,” Colman stated, citing latest high-stakes incidents such because the $25 million wire fraud in Hong Kong powered by a deepfake video and the deepfake robocall in New Hampshire supposedly from US President Joe Biden.
Colman stated his firm’s know-how makes use of an “ensemble strategy” to detect fraudulent audio, video, pictures, and textual content. The API, he stated, may be put in on-premises or within the cloud and be built-in into any utility. He cited Lexis-Nexis Threat Options’ prediction that deepfake-fueled fraud might value $1 trillion a 12 months. Whereas he did not go deeply into how the know-how really works, one of many slides confirmed an AI detector looped right into a voice name to find out whether or not the client voice is stay or manipulated.
“Our imaginative and prescient is to grow to be the detection layer for all AI-generated fraud,” Colman stated.
Judges requested about enterprise makes use of of the platform.
“So for banks, the No. 1 use is real-time voice fraud detection,” Colman answered. “All we do is say human or not human” — that’s, simply whether or not it is manipulated quite than attempting to establish the person.
When requested what was the corporate’s most difficult job, Colman stated, “Hiring nice PhDs specializing in engineering and analysis. And if anybody within the room is searching for a job, we’re hiring.”
The Remainder of the Finest
In the course of the Innovation Sandbox competitors, representatives from 10 cybersecurity startups had three minutes to pitch their know-how to a panel of judges after which confronted three minutes of questions from the judges. As in earlier years, the judges assessed finalists utilizing the next standards: the issue an organization units out to unravel and who they’re fixing it for; the originality and soundness of its mental property; the corporate’s go-to-market technique, its attain, and the energy of its workforce; and whether or not the strategy has been validated by the market — that’s, whether or not the corporate is earning money from actual prospects.
Along with Howe, this 12 months’s judging panel included Asheem Chandna, companion at Greylock; Dorit Dor, chief know-how officer of Examine Level Software program; Paul Kocher, unbiased researcher and founding father of Cryptography Analysis; and Nasrin Rezai, senior VP and CISO at Verizon.
Along with Actuality Defender, the different finalists this 12 months had been (in reverse alphabetical order): VulnCheck, RAD Safety, P0 Safety, Mitiga, Harmonic Safety, Dropzone AI, Bedrock Safety, Antimatter, and Aembit.
VulnCheck crawls the Web for CVE knowledge, then recombines and sanitizes it to create exploit and vulnerability intelligence for enterprise and product groups. When Kocher requested how they maintain from burning out, VulnCheck chief advertising and marketing officer Tom Bain credited prioritizing duties and giving groups knowledge early.
RAD Safety takes a behavioral strategy to cloud-native detection and response to handle evolving threats. RAD “fingerprints” the company’s setting so to evaluate new behaviors in opposition to the prevailing norms utilizing giant language fashions (LLMs).
P0 Safety provides a platform for cloud-access governance that secures all identities, whether or not human or machine. “If you need safe entry to the cloud, speak to us,” stated co-founder and CEO Shashwat Sehgal.
Mitiga sells a cloud investigation and response automation (CIRA) platform to supply safety operations facilities with cloud and software-as-a-service visibility. With extra assaults shifting to the cloud, the corporate will get calls “on a regular basis” as as to whether an organization has been breached, stated co-founder and CTO Ofer Maor.
Harmonic Safety helps organizations undertake generative AI whereas sustaining the safety and privateness of their knowledge, utilizing Harmonic’s prebuilt knowledge safety LLMs to detect delicate knowledge. Many of the workforce got here from Digital Shadows, co-founder and CEO Alastair Paterson’s earlier startup. ReliaQuest purchased Digital Shadows for $160 million in 2022.
Dropzone AI offered his firm’s AI SOC analyst that autonomously investigates alerts across the clock. When requested by the judges whether or not Dropzone was trying to exchange the human SOC analyst, founder and CEO Edward Wu stated, “We’re not trying to exchange the SOC analyst; we’re trying to increase the SOC analyst.”
Bedrock Safety makes an information safety platform powered by AI Reasoning Engine (AIR). “Fashionable enterprises are very similar to the Titanic on the subject of knowledge: They’re speeding by icy waters, unaware of regulatory icebergs and pirates which might be glad to ransom their knowledge,” stated co-founder and CEO Pranava Adduri.
Antimatter focuses on making knowledge protected to make use of in AI and machine studying purposes by mechanically classifying and redacting delicate knowledge, permitting customers to outline entry insurance policies, and implementing insurance policies irrespective of the place knowledge is saved.
“Antimatter is an information management airplane that sits between your purposes and your knowledge,” stated co-founder and CEO Andrew Krioukov.
Aembit sells a workload IAM platform that secures entry between workloads throughout clouds, SaaS, and knowledge facilities with out passwords. Firm co-founder and CEO David Goldschlag stated that IAM has secured person entry for 20 years, however there are 10 occasions extra workloads than customers.
