The research aimed to identify risks posed by GPAI models in specific cases (e.g., natural disasters, education, or health) and evaluate them according to the provision of the AI Act from a downstream provider perspective that recognises inherited risks. It is motivated by the recognised proliferation in the adoption of GPAI models for a variety of applications and the belief that to achieve responsible AI, we must first get better at evaluating it.
Aikajakso
11 lokak. 2024
Pidetty
Finnish Center for Artificial Intelligence FCAI, Suomi