AI Agents in the Wild: A Candid Conversation on Trust and Accountability - #NYTechWeek

Saturday, Jun 6

11:00am – 1:00pm

ETGMT

Hosted by

◔◟◔

⚆◟⚆

-‿-

+5

We’re deploying AI agents faster than we can verify them. As AI agents take on increasingly consequential roles, the question of trust has never been more urgent. What does it actually mean to trust AI agents and how do you build responsible infrastructure? Trust in AI agents is still largely improvised - based on intuition, brand, or retrospective debugging - rather than something measurable, inspectable, and enforced at the system level. We’re convening the best group of leading VCs, founders, academics, and product leaders in New York to explore a core question: what does it truly mean to trust AI systems - and what would it take to build a reliable trust layer around them? Through a focused panel and small-group roundtables, we’ll move beyond surface-level conversations and examine questions like: -How should we evaluate agents that plan, reason, and take multi-step actions? -What does reliability mean in systems that evolve across tool use, memory, and environment? -Where are the current blind spots in agent evaluation, monitoring, and control? -What signals - technical or behavioral - could make agent trustworthiness observable in real time? Grounded in lessons from high-stakes domains like finance and healthcare - where risk is modeled, audited, and continuously monitored - we’ll explore how similar principles might translate (or fail to translate) to modern AI systems. Topics will include hallucination and bias detection, evaluation blind spots, human-in-the-loop design, and emerging approaches to quantifying model reliability. This is more than a conversation about “responsible AI” as a principle. It is a working session on something more concrete: how to move from intuition-based trust to system-level guarantees - and what needs to exist, technically and institutionally, for that shift to happen.

Guest List

2 Going · 1 Interested

View all

˜▾˜

*⏖*

❛ᴗ❛