TELUS Digital Launches Fuel iX™ Fortify for Automated Red-Teaming

TELUS Digital , a leading global technology company specializing in digital customer experiences (CX), announced the launch of its continuous automated red-teaming application, Fuel iX™ Fortify. The solution helps enterprises test generative AI (GenAI) systems at scale and identify vulnerabilities by replicating real-world attack scenarios through advanced adversarial simulation techniques.

Also Read: AiThority Interview with Tim Morrs, CEO at SpeakUp

Concerns about safety and security vulnerabilities are a critical consideration for organizations adopting GenAI, yet over two-thirds of enterprises (67%) report a shortage of cybersecurity professionals to address them. This challenge highlights the importance of proactive safeguards that can help security teams keep pace with rapid AI development. Fuel iX Fortify runs thousands of adversarial attacks on AI-powered assistants, copilots and applications in minutes, helping organizations uncover AI risks before they can be exploited. With an ever-evolving database of adversary tactics, techniques and procedures based on the latest research, Fortify stays ahead of emerging attack methods and generates novel tests for each safety assessment, or it can rerun specific attack formulations to track over time. The Gen AI red-teaming app can also automatically create unique attack objectives tailored to a system’s code of conduct policy, making it purpose-built for the unpredictability of GenAI, which can generate different answers each time it is asked a question.

“GenAI is driving enterprise transformation, but it’s simultaneously creating new safety and security risks, that without safeguards, can erode customer trust, damage brand reputation, disrupt operations and expose companies to regulatory penalties,” said Bret Kinsella, General Manager, Fuel iX at TELUS Digital. “Red-teaming expertise is in short supply, and many organizations struggle to find the people and time needed to keep pace with evolving risks. Fuel iX Fortify helps close that gap by automating AI safety and security testing and uncovering vulnerabilities early so security teams can proactively mitigate risk. Importantly, it is accessible to non-technical users and backed by research. Fortify was built as an AI safety and security early warning system that is available to identify risks at any time by any user to give enterprises confidence they are responsibly launching and scaling their AI applications.”

Enterprise-ready features of Fuel iX Fortify

Fuel iX Fortify supports enterprises across all industries with AI security features that include:

• Accessible design: Intuitive interface enables both technical and non-technical teams to conduct AI security testing.
• Actionable reporting: Dashboards and detailed reports provide users with a clear overview of each AI application’s vulnerabilities, making it simple to understand and act on security and safety risks.
• Seamless integration: Can be used with a wide array of GenAI applications, simultaneously within a single interface.
• Advanced novel and repeat attack generation: Stays current with emerging threats and attack methods through regular updates based on the latest research.

Applying Fuel iX Fortify in healthcare

Fuel iX Fortify is already in use by a publicly-funded health authority in Canada to safeguard its public-facing career advisor chatbot. The bot provides a personalized, user-friendly experience, helping individuals explore job opportunities in the healthcare sector in the same way they would interact with a recruiter, ensuring it would never provide medical advice or harmful responses. With Fuel iX Fortify, the client experienced a 97% reduction in testing time, while it took less than 30 minutes to onboard non-technical experts.

The client stated:

“Fuel iX Fortify gave us confidence that our chatbot’s responses were safe, appropriate and met our standards. We found that manual red-teaming was slow, limited in scope and produced error rates as high as 13%. Fortify let us test our chatbot thoroughly before launch, so we knew it wouldn’t give harmful answers or fall short on compliance. With that assurance, we were able to launch responsibly, protect our reputation, and create a positive, human-like experience for people interested in healthcare careers.”

Also Read: Cognitive Product Design: Empowering Non-Technical Users Through Natural Language Interaction With AI-Native PLM