목차
Title page 1
Contents 1
Overview 3
Scope of risks 3
Current state of the science 4
Research priorities 7
1. How can safeguards built into AI models and systems to reduce harmful outputs be evaluated and improved? 7
2. How can current DCT techniques and their implementations be evaluated? 8
3. How is synthetic content spreading in the information environment and what systemic impacts is it having (e.g., with regard to public trust)?... 11
4. What technical or non-technical approaches can be used to support or advance DCT techniques? 12
Glossary 15
