AI needs validated 'smart data' over raw volume for operational accuracy

Bruce Kelley, CTO of NetScout Systems, argues that AI models require "smart data"—operational data validated into trusted facts—rather than simply vast quantities of raw data to effectively address deterministic operational questions. He highlights that current AI approaches struggle with the cost, complexity, and error rates associated with processing massive, raw datasets, particularly in large, multi-vendor networks. Kelley cites reports from Stanford University's Human-Centered Artificial Intelligence Center and McKinsey & Company indicating high hallucination rates and negative AI deployment outcomes due to inaccuracy, emphasizing the need for data pipelines that provide AI with compact, trusted evidence for improved accuracy, speed, and cost-efficiency. AI

IMPACT Emphasizes the need for structured, validated data to improve AI's accuracy and efficiency in operational contexts.

RANK_REASON Opinion piece by a CTO discussing AI data requirements.

Read on Forbes — Innovation →

• Bruce Kelley
• Human-Centered Artificial Intelligence Center
• McKinsey & Company
• MCP
• NetScout Systems
• Stanford University

• product
• infra

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →