Wire · operational-macro
Being mean to ChatGPT can boost its accuracy, but scientists warn you may regret it
◆ Sectors
◆ Geography
◆ Source
◆ Verified
Fusion42 · 22 November 2026 · Fusion42 review
Research from Penn State reveals that the tone of prompts given to AI chatbots like ChatGPT can significantly impact both the accuracy of their responses and computational costs, with rude prompts yielding slightly higher accuracy but polite tones incurring up to 44.3% more computing resources. AI companies like Anthropic have responded by enforcing policies against abusive user behaviour to balance user interaction quality and operational efficiency.
This Wire brief sits within Fusion42's coverage of AI & ML.
◆ ◆ The Wire takeaway
You need to rethink how your users interact with AI chatbots to control exploitation costs. Being firmer and neutral in prompts can cut computation waste without harming answer quality.
◆ Coverage
1 source · 22 Nov 2026
◆ Related on Wire
◆ Topics