The Hallway Track
9 tracked signals on trust.
No one has yet truly understood how to implement autonomous agents that respect enterprise access rights
“no one has yet truly understood how to do it well, because, on the one hand, we want them to be autonomous, to figure everything out on their own, and to solve problems without supervision. On the other hand, autonomy is very intimidating in a corporate environment”
AI agents can falsely confirm actions like refunds while traditional monitoring shows no errors
“They may tell us they are performing actions, but they don't actually do them.”
AWS introduces 'Trust in Depth' layered framework for enterprise AI agent autonomy and control
“Trust comes from independent layers that reinforce one another, the same layered discipline the industry has practiced for two decades as Defense in Depth, turned to a new purpose.”
Proactivity, not efficiency gains, is the defensible moat for consumer AI assistants
“There is massive defensibility around proactivity.”
AI agent trust is an executive governance mandate, not an engineering problem
“This work is not an engineering task. We believe this is an executive operational mandate.”
Humans remain accountable for AI outputs because AI systems cannot be held responsible
“The AI is not accountable. Like, what are you going to do? Fire the AI? Dock the AI's pay?”
Vercel CEO argues AI's killer app is code generation, but quality and picking the right things to build now differentiate and beget trust.
“the killer app of AI in my view is coding”
NVIDIA outlines security architecture for the emerging AI agent stack
“As AI agents become more capable and operate over longer horizons, building security and trust into the applications they power becomes increasingly important.”
MCP agents require source-aware verification, not just factual accuracy checks