look news india
News India Live, Digital Desk: Whenever we ask something from ChatGPT or any other AI chatbot, we often write “Please, will you write this email for me?” Or don’t forget to say “Thank you” when you get a reply. We believe that if we speak well, AI will also work well.
But, a recent new study has turned this entire theory upside down. Scientists and data experts say that if you want accurate and to-the-point answers from ChatGPT, stop saying “please” and “thank you.” yes, a little ‘Rude’ Or being strict may be beneficial for you.
Why doesn’t AI understand the language of love?
It sounds a bit strange that how can work be improved by misbehavior? Actually, there is a technical reason behind this.
When we use a lot of words like “Would you please…”, “Kindly…” or “I was wondering if…”, we get a lot of feedback in the prompt. ‘Noise’ Let’s fill it. AI model is a machine. When he receives direct and strict instructions, it is easier for him to understand what you want.
Too much polite language sometimes confuses the AI and it deviates from the main issue.
What came out in the study?
Research showed that people who “Without any hesitation” Given instructions in direct and strict words, they got more accuracy.
- Polite approach: “Can you please tell me how to do this?” (In this the AI has to think more).
- Strict/rough approach: “Explain it to me. Now.” (This is a neat command for AI).
The scientists found that ‘rude’ or ‘direct’ prompts provided better answers to complex questions in many cases. It seems that when we apply pressure or give direct orders, the AI focuses its entire ‘mind’.
So should we now resort to abuse?
Oh no! This does not mean that you should start criticizing AI. This simply means that ‘Professional’ Become.
Give commands just like you would give orders to a junior. Do not write useless things. Your time will be saved and you will get good answers also.
So next time you open ChatGPT, try talking directly about work instead of ‘please’, you might notice the difference!
