Op-ed: AI risk cannot be governed without independent researcher data access

What the research found

Op-ed from Tech Policy Press arguing AI risk cannot be governed without independent researcher access to data, drawing on social media research lessons and a Princeton Accelerator analysis of 113 researcher requests.

Claims and how well the evidence supports them

1. Researchers face industry opacity that limits independent assessment of AI risk 98%
2. Princeton Accelerator analyzed 113 unique researcher requests 94%
3. Digital trace data has answered questions on trafficking, propaganda and drug markets 98%

Evidence

• techpolicy.press https://techpolicy.press/you-cant-solve-ai-risk-without-solving-the-ai-evidence-problem

Actionability rated none.