Protect AI, an innovative startup dedicated to developing solutions for bolstering AI systems' security, disclosed today that it successfully garnered $35 million in a Series A funding round. The round was spearheaded by Evolution Equity Partners and witnessed active participation from Salesforce Ventures, Acrew Capital, boldstart ventures, Knollwood Capital, and Pelion Ventures.
The value of this investment round is over double the amount raised in Protect AI's seed round, concluded last December, collectively inflating the startup's total funds to a staggering $48.5 million. According to co-founder and CEO Ian Swanson, the newly raised capital will be channeled into growing Protect AI's platform capabilities, broadening their research scope, and kickstarting fresh open source projects.
Swanson communicated to TechCrunch via email, “Now, we have plenty of capital to weather the storm for years to come,” segueing into the startup's employment expansion plans from its present 25-member team to around 40 employees by the close of this year.
Swanson and Daryan Dehghanpisheh became the torchbearers of Protect AI, introduced to the world in 2022. Individually, their professional journey involves illustrious tenures at AWS and Oracle, contributory roles in launching DataScience.com, an AI development platform later procured by Oracle, their past employer.
In Swanson's words, "We conceived Protect AI 18 months prior, riding on our expertise being a part of the major machine learning and AI deployments around the globe. Our sights were set on the immense potential AI offered, but they didn't sway from the inherent perils these systems housed. Our mission twirled around aiding our patrons in crafting a safer AI-driven world."
The product portfolio of Protect AI includes AI Radar alongside tools adept at warding off certain types of AI-centered attacks, for example, prompt injection onslaughts. Prompt injection alludes to the circumstance where text-based instruction-operated AI, known as prompts, are duped by adversarial prompts into executing tasks straying from their original aim.
Additionally, Protect AI furnishes utilities that can scrutinize documents from Jupyter Notebook, a popularly utilized platform for AI model formulation and data science experimentation. Research companies have discovered that Jupyter Notebook files that lack adequate security can fall prey to Python-based ransomware and cryptocurrency mining attacks.
Protect AI conscientiously checks Jupyter notebooks for various potential issues—personally identifiable data (e.g., names and phone numbers), proprietary authentication tokens and credentials, and open source code that carries a “nonpermissive” license capable of barring its application in commercial systems.
Swanson further elaborated on the dire necessity of transitioning from machine learning operations (MLOps), a time-tested practice spanning over a decade for firms, to embracing security. He urged, "We must reach a point where we seamlessly undertake ML security operations — 'MLSecOps' — on a large scale within vast enterprises."
It's clear that the adoption of robust security measures in AI-powered systems is becoming more critical than ever. Solutions like those offered by Protect AI play a pivotal role in this changing landscape, ensuring a balance between innovation and security. As more businesses turn to AI to drive efficiency and productivity, platforms like this will be invaluable in safeguarding against potential threats.
Like AppMaster, Protect AI aims to simplify and secure AI operations for the benefit of businesses worldwide. This shared ethos showcases the increasing significance of no-code in various niches, from app development to AI security, changing the digital landscape for businesses worldwide.