Algorithmic Accountability
Who is responsible when an algorithm makes a harmful decision, and how transparency and auditing can create accountability.
When Algorithms Decide
Algorithms increasingly make consequential decisions: who sees which news, who gets approved for a mortgage, which neighborhoods are policed most heavily, which resumes reach human reviewers, and which patients are prioritized for treatment. When these systems produce harmful outcomes, identifying who is responsible is extraordinarily difficult. The developer who wrote the code, the company that deployed it, the government that required it, and the data that trained it all share some responsibility, but existing legal frameworks struggle to assign it.
This is the algorithmic accountability gap. Traditional liability requires showing that a person or entity made a decision that caused harm. But when an algorithm makes a decision, no individual chose the outcome. The system may have been designed with good intentions but trained on biased data, or it may work perfectly in aggregate but produce unjust results for specific individuals or communities.
