Businesses are reducing their artificial intelligence budgets after major providers introduced per-token pricing models. This shift marks the end of the era where organisations could generate unlimited text without worrying about direct costs. The expense structure now forces teams to reconsider how they deploy large language models for routine tasks. Some firms are experimenting with tools to restrict model outputs to basic vocabulary, effectively forcing the systems to communicate like non-native speakers. This tactic aims to bypass complex processing charges while retaining basic functionality. Meanwhile, the market faces new fraud risks as scammers sell seeds for exotic flowers that exist only in digital images. These fake products appear on major platforms like eBay, Etsy, and Amazon, exploiting the confusion around synthetic media. The trend highlights a transition from unchecked growth to careful cost management in enterprise technology.
- Providers now charge fees based on the number of tokens generated
- Companies are limiting model vocabulary to reduce processing costs
- Fake AI-generated flowers are being sold as real products online
