This week we spotlight the tenth risk framework included in the AI Risk Repository:
This study aimed to systematically review the main social impacts of Artificial Intelligence, as well as the most common strategies to mitigate impacts.
In total, the authors included and extracted impacts from 175 articles, identifying 9 main categories of social impact from AI:
These categories were determined according to themes that were most frequently mentioned in the included articles. The three most mentioned categories of social impact were Bias and Discrimination (26%), Risk of Injury (16%) and Data Breach/Privacy and Liberty (15%).
Reviews and summarises social impacts identified in the AI literature based on how frequently they are discussed
Reviews and summarises recommended mitigation strategies from the AI literature based on how frequently they are discussed
Classifies included papers according to the domain of AI use (e.g., health, education, society, military etc.)
This summary highlights a paper included in the MIT AI Risk Repository. We did not author the paper and credit goes to Paes, V. M., Silveira, F. F., Akkari, A. C. S. (2022). For the full details, please refer to the original publication: https://doi.org/10.1007/978-3-031-04435-9_54
→ View all the frameworks included in the AI Risk Repository
→ Explore the AI Risk Repository Website