Toward User-Driven Algorithm Auditing: Investigating users’ strategies for uncovering harmful algorithmic behavior
A three-stage model capturing the process by which people uncover biased and harmful algorithmic behaviors: Search Inspiration, Sense Making and Remediation
- Search inspiration: encompasses ways in which users come up with ideas for where to look for potential biases in an algorithmic system
- Sense making involves the ways that users understood and evaluated algorithmic behaviors that might be harmfully biased, and
- Remediation: participant actions and desires to mitigate the ramifications of harmful algorithmic bias.