Experts say that air-gapping AI could prevent events like the Hugging Face hack, but would undermine the value of evaluations and slow research to a crawl (Robert Hart/The Verge)

Experts suggest that air-gapping AI could prevent incidents like the Hugging Face hack. However, this approach would significantly undermine the value of evaluations and drastically slow down research progress. Researchers are currently testing these systems specifically because of their potential for unpredictable or even dangerous behaviors, indicating a need for continued, accessible evaluation.

This report uniquely highlights the trade-off between AI security measures like air-gapping and the practical necessity of open evaluation for research, unlike other discussions focusing solely on security or research benefits.