Jump to Content

Prompting Considered Harmful

Published
View publication Download

Abstract

As a computer scientist with one foot in AI research and the other in HCI research, I have become increasingly concerned that “prompting” has transitioned from what was essentially a test and debugging interface for ML engineers into the de-facto interaction paradigm for end-users of large language models and their multimodal generative AI brethren. It is my professional opinion that prompting is a poor user interface for generative AI systems, and should be phased out as quickly as possible.

My concerns about prompting are twofold. First, prompt-based interfaces are confusing and non-optimal for end-users (and ought not to be conflated with true natural language interactions). Second, prompt-based interfaces are also risky for AI experts – we risk building a body of apps and research atop a shaky foundation of prompt engineering. I will discuss each of these issues in turn, below.

Authors

Meredith Ringel Morris

Venue

Communications of the ACM