# A Framework for Understanding Sources of Harm throughout the Machine Learning Life Cycle

## TL;DR

This paper presents a comprehensive framework identifying seven sources of harm in the machine learning life cycle to improve understanding, communication, and mitigation of potential negative societal impacts.

## Contribution

It introduces a novel framework that systematically categorizes sources of harm across data collection, development, and deployment stages of ML.

## Key findings

- Identifies seven distinct sources of downstream harm in ML.
- Facilitates better communication about harm sources.
- Supports targeted mitigation strategies.

## Abstract

As machine learning (ML) increasingly affects people and society, awareness of its potential unwanted consequences has also grown. To anticipate, prevent, and mitigate undesirable downstream consequences, it is critical that we understand when and how harm might be introduced throughout the ML life cycle. In this paper, we provide a framework that identifies seven distinct potential sources of downstream harm in machine learning, spanning data collection, development, and deployment. In doing so, we aim to facilitate more productive and precise communication around these issues, as well as more direct, application-grounded ways to mitigate them.