Five AI safety sessions every founder should have on their TechCrunch Disrupt 2026 agenda

Five AI safety sessions are highlighted for founders attending TechCrunch Disrupt 2026. These sessions cover various aspects of ensuring trust and security in AI systems. The first session, "What Anthropic Sees When Enterprises Actually Deploy Claude," explores the difference between successful enterprise deployments and stalled ones.

The second session, "The Agent Security Problem Nobody Is Talking About," focuses on securing AI agents at the infrastructure level, addressing application-level permission models and architectural decisions. The third session, "Securing the AI Enterprise: Why the Cloud Got a Lot More Complicated," discusses the need for security, governance, and observability in deploying AI in critical systems.

The fourth session, "Building AI Systems When Failure Is Not an Option," brings together perspectives from defense and autonomous systems industries to address creating a safety culture, testing, validating AI, navigating regulatory hurdles, and building trust when the consequences of failure are physical. Lastly, the fifth session, "Robots Are Waiting for Their ChatGPT Moment.

Here Is What Is Standing in the Way," explores data pipelines, simulation environments, and foundation models that could help close the data gap in physical AI, enabling the development of reliable and capable robots. These sessions will be held across the AI Stage and Real World AI Stage at Disrupt, along with 200+ additional sessions, making it a valuable opportunity for founders to gain insights into building trustworthy AI systems.