Navigating the Jagged Technological Frontier | Harvard Business School AI Institute

ai-productivityknowledge-workhuman-ai-collaborationjagged-frontierbcgchatgpt-4

Abstraction: Field experiment showing AI boosts consultant productivity unevenly across tasks

Key points:

Field experiment with 758 BCG consultants (7% of individual contributor workforce) testing ChatGPT-4 on real consulting tasks spanning creativity, analytical thinking, writing, and persuasiveness
Within-frontier tasks: AI boosted speed by over 25%, human-rated quality by over 40%, and task completion by over 12%
Introduces the "jagged technological frontier" concept: AI excels dramatically at some tasks while falling short on others, with the boundary non-obvious to users
Two distinct usage patterns emerged: "Centaurs" (humans and AI divide tasks between themselves) and "Cyborgs" (humans integrate AI into their full workflow)
Key argument: debate should shift from "adopt vs. not adopt AI" to evaluating optimal human-AI configurations per task type within knowledge workflows
Led by HBS chair Karim Lakhani in collaboration with Boston Consulting Group

Connections: Harvard Business School · Boston Consulting Group · Chatgpt · Human AI Collaboration · Knowledge Work · Large Language Models