Senior Software Architect At Okta Inc. San Francisco, USA

Title of the Talk:

Human-AI Cooperative Authentication: Beyond Passwords and CAPTCHA

Abstract of the Talk:

Authentication remains one of the most persistent and fragile foundations of digital trust. Despite decades of innovation, widely deployed mechanisms such as passwords, one-time codes, and CAPTCHA systems continue to impose high friction on legitimate users while offering diminishing resistance against automated, AI-driven attacks. As attackers increasingly leverage machine intelligence, static and challenge-based authentication models are reaching their practical limits.
This keynote argues that the future of authentication lies not in stronger challenges, but in cooperative interaction between humans and AI systems. Rather than treating authentication as a single verification event, human-AI cooperative authentication reframes it as a dynamic dialogue, where AI continuously evaluates contextual signals such as identity history, behavior, device posture, and intent to adapt authentication requirements in real time. Legitimate users experience reduced friction, while adversaries face escalating, personalized challenges that are difficult to automate or replay.
The talk explores how AI can act as a trust mediator rather than a replacement for human agency, determining when explicit human interaction is necessary and when confidence is sufficient to proceed transparently. It examines design principles, system architectures, and failure modes associated with this approach, including explainability, bias, adversarial probing, and usability trade-offs.
By moving beyond passwords and CAPTCHA toward cooperative, confidence-aware authentication, this keynote outlines a path toward systems that are not only more secure, but also more humane, balancing usability, resilience, and trust in an era of intelligent adversaries.