Perceived accuracy of automated risk assessment tools
Full Question Text
Q26. In some places, automated tools are used to make decisions about public safety risks that in the past were made by humans. For example, software tracking and keywords may be used to monitor social media for events where a crime might happen. In your opinion, are the decisions made by these tools (such as AI and machine learning) generally more or less accurate than those made by humans?
| Total | More accurate | 12% | About the same | 20% | Less accurate | 9% | Don't know | 59% |
| Office | Department | Total | More accurate | 17% | 11% | 12% | About the same | 16% | 21% | 20% | Less accurate | 16% | 8% | 9% | Don't know | 51% | 60% | 59% |