Scientists Call for Oversight and Guidelines to Ensure Responsible Development of AI like ChatGPT

Rapid advances in generative AI like ChatGPT require oversight to manage risks like misinformation and diminish validity of research.

Scientists should take the lead in testing and improving AI safety through an independent auditing body and living guidelines.

The guidelines advocate human verification of key research steps and transparency about AI use.

The auditing body would develop benchmarks to certify AI systems for accuracy, bias, and ethical issues.

Obtaining international funding and collaboration with tech companies is crucial to implement the guidelines and auditing body.