Ai policies help and hinder software developers in practice
AI Policies: Help or Hindrance? A Software Developer's Perspective
Software EngineeringArtificial Intelligence
Summary
Software organizations create rules to keep AI tools safe and private, but these rules don't work well if software developers don't follow them. The authors talked to 19 developers to learn how these AI policies help and also get in the way of their work. They found that policies need to fit developers' needs better to be useful. The authors suggest ways for managers to introduce AI policies that developers will actually use.
What this means in practice
- •For software engineering managers: Design AI policies that consider developer needs to improve compliance and reduce friction when using AI tools.
- •For enterprise security teams: Develop risk mitigation strategies that balance security concerns with developers’ workflows to ensure effective AI policy enforcement.
Authors
Samuel Ferino, Rashina Hoda, John Grundy, Christoph Treude, Hashini Gunatilake
Abstract
AI policies introduced by software organisations to mitigate LLM-related risks such as sensitive information leaks and unauthorised usage are not useful if software developers do not engage with them. We draw on 19 software developer interviews to show how AI policies help and hinder developers. We suggest approaches to support managers and decision makers with a developer-centric approach to introducing AI policies.