“We’re excited to introduce this much needed framework to the industry as we continue to help global changemakers build a better world by advancing responsible AI.”, says Phil Munz, Director of Data Science at TrojAI.
SAINT JOHN, New Brunswick (PRWEB) April 20, 2022
TrojAI Inc. announced its publication of the industry’s first Secure Model Development Life Cycle (SMDLC). Just like cybersecurity permeated into traditional software engineering with the Secure Software Development Life Cycle (SSDLC), TrojAI has recognized and filled a similar need with AI/ML model development by ‘shifting cybersecurity left’ and defining best practices to be followed throughout the model development process.
“Increasingly, AI/ML models are becoming core to enterprise systems and the cybersecurity of those models has become a priority for the safe and secure deployment of responsible AI”, says James Stewart, founder and CEO of TrojAI, “and we’ve realized that our clients needed a formal framework to assist them in their efforts to deploy trustworthy AI.”
SMDLC outlines the many threat vulnerabilities that need to be considered from Problem Identification & Design through to Deployment & Monitoring of models and recommends how best to minimize such dangers as data poisoning and model evasion attacks due to supply chain and other points of vulnerability. Given this broader context, an organization can improve its overall security posture to help ensure the deployment of safe and responsible AI.
The release of the SMDLC is also timely given the looming regulations around responsible AI that can cost organizations, in some cases, upwards of 6% of global revenue for failing to protect AI/ML systems. “We’re excited to introduce this much needed framework to the industry as we continue to help global changemakers build a better world by advancing responsible AI.”, says Phil Munz, Director of Data Science at TrojAI.
Since 2019, TrojAI has been developing cybersecurity tools focused on improving both the robustness and security of AI models through data audits, model assessments and AI firewalls for computer vision, natural language processing and tabular/structured models. Over the past year, TrojAI Inc. has participated in the Techstars Montreal AI accelerator, the Rogers Cyber Catalyst program out of Ryerson University, and is currently working with both ventureLAB and Creative Destruction Labs, in Toronto, Canada. TrojAI is supported by the National Research Council of Canada and is a Gartner Representative Vendor for Adversarial Resistance.
Those interested in deploying robust and secure AI/ML models can request their own copy of the SMDLC framework.
- TrojAI Inc. is a Canadian-based firm dedicated to building cybersecurity solutions to enhance and protect AI innovations from both naturally occurring and adversarial data poisoning and model evasion attacks which can adversely influence model behaviours.