AINeutralarXiv – CS AI · 10h ago6/10
🧠
Can LLMs Reason About Brand Ownership? An Empirical Study of Domain Attribution Intelligence
Researchers empirically evaluated whether large language models can reliably determine domain ownership for brand protection purposes. The study found that while LLMs achieve 82% precision enumerating brand domains from memory, they fail at ownership verification without external tools (F1 score of 0.37), but WHOIS augmentation dramatically improves performance to near-perfect precision, reducing false positives that harm users and brand reputation.
🧠 Claude🧠 Sonnet🧠 Gemini