Two Yale PhDs Are Trying to Make AI Hallucinate 10x Less, So Small Businesses Can Adopt the Technology in Weeks, Not Years Enkrypt AI announced a $2.35 million funding round on Tuesday.
By Sherin Shibu •
Key Takeaways
- AI can get things wrong and hallucinate — usually at a rate of 6%.
- Two Yale PhDs say they can bring that rate down to 0.6%, and they want to fast track AI adoption in small businesses from 1-2 years to weeks.
- They claim to have the only platform that combines visibility and security for Gen AI on a business level.
Enkrypt AI, a startup focused on making generative AI and large language models (LLM), is trying to make AI hallucinate 10 times less. It claims that its platform can help small businesses and larger enterprises adopt AI in weeks, not years.
The company announced a $2.35 million funding round on Tuesday.
Yale PhDs Sahil Agarwal (CEO) and Prashanth Harshangi (CTO) cofounded Enkrypt AI in 2022 after conversations with executive leaders convinced them businesses needed to trust AI first before they could adopt the technology. AI models have been known to hallucinate or churn out incorrect results, at a rate of 3% to 10% or even higher, which makes them risky bets for businesses.
Agarwal and Harshangi claim that Enkrypt AI's platform is the only one out there that makes gen AI apps secure and visible on an enterprise level. They say they have reduced hallucinations and vulnerabilities with AI and LLMs from 6% to 0.6%, a tenfold decrease, so that enterprises can confidently adopt AI without delays.
"Based on our conversations with CIOs, CISOs and CTOs, we are convinced that for LLMs to be widely adopted, it must be built on a foundation of security, privacy, and compliance," Agarwal stated in a press release.
Enkrypt AI co-founders and Yale PhDs Sahil Agarwal and Prashanth Harshangi
Enkrypt AI's platform aims to add layers of safety and security between users and AI models and to make everything from chatbots to automated reports more efficient. The team has worked on creating and deploying AI models before, including at the U.S. Department of Defense.
Related: AI Is Coming For Your Jobs — Anyone Who Says Otherwise Is In Denial. Here's How You Can Embrace AI to Avoid Being Left Behind.Enkrypt AI's platform does more than detect vulnerabilities, according to Harshangi; it gives in-house developers at companies tools to protect their AI solutions.
"As the benefits of AI become ever more tangible, so do the risks," Harshangi stated.
Related: Why We Must Focus on What AI Can Give Us Instead of Focusing on What It Could Take AwayA Fortune 500 data infrastructure company is using Enkrypt AI now to detect and resolve LLM attacks and prevent data leaks, the company said.
The $2.35 million funding round was led by Boldcap, with participation from Berkeley SkyDeck, Kubera VC, Arka VC, Veredas Partners, Builders Fund, and angel investors.
