GenAI: An Unreliable Information Store - Enhancing LLM Trustworthiness in Enterprise Environments

Explore the challenges and solutions of integrating Large Language Models (LLMs) into enterprise environments in this insightful conference talk by Noble Ackerson. Delve into the inherent unreliability of LLMs and discover innovative approaches to enhance their trustworthiness in critical applications. Learn about the use of vector databases and retrieval augmented generation techniques. Gain valuable insights from Noble's extensive experience as an AI Product Manager, Google Design Sprint Master, and former Google Developers Expert. Understand the complexities of data trust, product output, and the cooperative principle in AI integration. Examine the concept of overfitting and its drawbacks, as well as the complexity paradox in LLM implementation. Explore methods for calibrating trust and facilitating effective conversations with AI systems in enterprise settings.

Introduction
The Common Thread
Data Trust
Field Report
Product Output
Complexity
The Cooperative Principle
The Complexity Paradox
Overfitting
Drawbacks
Complexity Paradox
Summary
Calibration of trust
Facilitating conversation