1 paper
Kamilė Lukošiūtė, John Halstead, Luca Righetti
AI companies and governments are increasingly concerned about frontier AI systems enabling cybercrime, yet defining meaningful capability thresholds requires knowing the scale of c…