Google’s AI chatbot, Gemini, is facing “distillation attacks” from commercially motivated actors attempting to clone its capabilities through extensive querying, including a single campaign with over 100,000 prompts. These attacks aim to expose the underlying patterns of the model, which Google considers intellectual property theft. The company believes the attackers are primarily private companies or researchers seeking competitive advantages. As AI development intensifies, similar vulnerabilities are expected to affect smaller companies’ custom AI tools.
Want More Context? 🔎
Loading PerspectiveSplit analysis...
