Misinformation/Disinformation & Trust

AI is dominating everything—from attacks over the stock market to platforms addressing Zero Trust. It seems like an overnight success, but looking back to 1967, when Marvin Minsky predicted that “Within a generation... the problem of creating 'artificial intelligence' will be substantially solved,” it becomes clear that it was a long way to get to where we are now.

AI simplifies the processing of vast amounts of data and helps us draw meaningful conclusions. We have learned that humans make mistakes, and it is our experience that computers always give us the correct response when used for data calculations. AI seems to make computers more similar to humans—more intelligent but also more error-prone.

Unfortunately, we have an asymmetry in cybersecurity. With millions of attacks each day, we need high protection success rates—even a 99% success rate would still mean 10,000 successful attacks. In contrast, attackers only need a success rate of 30% to achieve their goals, creating a significant advantage for "AI for bad."

In the era of passkeys, how do we balance the robust security primitives, based on provably secure protocols, with probabilistic artificial intelligence to respond to that asymmetry?

In the era of generative AI, misinformation can be produced faster and more convincingly than ever before—and often with plausible-sounding digital identities. As deepfakes, synthetic personas, and manipulated content multiply, how can we verify who really said what, and ensure digital trust without undermining privacy or stifling innovation?

This panel explores the intersection of identity security and content authenticity, asking critical questions about attribution, accountability, and assurance in a world where machines can imitate humans and humans can exploit machines.