Many companies ship AI chatbots as thin wrappers around commercial LLM APIs with zero cost controls. What if a tool behaved like an overly engaged, perfectly valid user - and just burned through their budget?
For every $x of increased revenue that the AI company gets from their customer whose chatbot is being abused you would have to also use $x of tokens from the free trials of that same AI company.
Cool idea however, I don’t want to benefit AI companies in the process.
Wasting the energy isn’t that good either…
Probably would create a demand and make their numbers go up.
Sure they lose money at the end of the day but this could delay their ai bubble burst.
For every $x of increased revenue that the AI company gets from their customer whose chatbot is being abused you would have to also use $x of tokens from the free trials of that same AI company.