
Bret Kinsella is VP of AI Safety and Security at TELUS Digital, where he helps enterprises move from AI guardrail guesswork to systematic, scalable red teaming that keeps generative AI both safe and useful in production environments.
Bret Kinsella, VP of AI Safety and Security at TELUS Digital, is helping enterprises move from AI guardrail guesswork to systematic, scalable AI red teaming. TELUS Digital works with some of the world's largest companies to test, evaluate, and harden AI systems before they reach customers, using structured adversarial testing to find failure modes that standard QA processes miss entirely.
In this episode, Russ and Bret break down what AI red teaming actually involves, why most enterprise AI deployments are more vulnerable than their builders realize, how TELUS Digital runs red team exercises at scale across different industries and use cases, and what it takes to build a genuine safety and security function around AI systems that are constantly evolving.
They also get into the difference between AI safety and AI security, why adversarial testing needs to be continuous rather than a one-time pre-launch exercise, how Bret thinks about the emerging regulatory landscape around AI risk, and why the companies that take AI safety seriously now will be the ones that earn long-term customer trust and avoid the incidents that set entire industries back.
Topics Covered:
- Welcome and intro; Bret Kinsella and TELUS Digital AI safety work
- What AI red teaming is and how it differs from standard QA
- Why most enterprise AI deployments have more vulnerabilities than expected
- How TELUS Digital runs red team exercises at scale
- Testing AI systems across different industries and use cases
- The difference between AI safety and AI security
- Why adversarial testing needs to be continuous not just pre-launch
- Building a genuine safety and security function around evolving AI systems
- The emerging regulatory landscape around AI risk
- Why taking AI safety seriously now builds long-term customer trust









