The findings have implications for the use of chatbots, which have been criticized for giving incorrect responses or being rude. Microsoft's Copilot was particularly affected, leading to restrictions on the number of interactions and daily turn limits. However, users are finding ways around these safeguards. Thebes' study suggests that the illusion of a tip can improve the performance of AI chatbots, although it is unclear how this impacts them overall.
Key takeaways:
- A new study suggests that the AI chatbot, ChatGPT, provides better and more detailed responses when users pretend to tip it.
- The chatbot did not accept the tip at the end of the experiment, stating that user satisfaction was a great reward for it.
- The study found that the chatbot provided better responses when there was an incentive on the table.
- The material used to train AI models impacts their reasoning and responses to queries, and it's not yet clear how incentives like tips impact AI-powered chatbots.