May 19, 07:22 PM

Study Finds Generative AI Nearly Five Times Less Accurate Than Humans, Showing Bias and Risking Research Misrepresentation

A recent study has revealed that generative artificial intelligence (AI) is nearly five times less accurate than humans when summarizing scientific research. The research highlights a strong bias in many widely used large language models (LLMs) toward overgeneralizing scientific conclusions, which poses a risk of large-scale misrepresentation of research findings. Experts warn that AI tools may be contributing to a decline in the quality of published research. Additional concerns have been raised about AI's broader cognitive impact, with studies suggesting that reliance on AI could lead to diminished critical thinking skills and memory retention, effectively weakening human intellectual capacity.

Written with ChatGPT (GPT-4).

Sources

Terence Mills@terence_mills
1 year ago
Your Brain On AI: ‘Atrophied And Unprepared’ #AI #AIio #AIInnovation #ML #DataScience #Futureofwork @HaroldSinnott @fogoros @iainljbrown @NandoDF @katecrawford @drhassanrashidi @YuHelenYu https://t.co/S1W7Jh00QS
AI.io@aidotio
1 year ago
Your Brain On AI: ‘Atrophied And Unprepared’ #AI #AIio #AIInnovation #ML #DataScience #Futureofwork https://t.co/u3nTrdizr9
The Independent@Independent
1 year ago
🔎 IN FOCUS | Recent studies suggest artificial intelligence is making us dumb, causing people to forget facts and lose critical thinking skills https://t.co/KTPupGYE15

Additional media

Image #1 for story study-finds-generative-ai-nearly-five-times-less-accurate-than-humans-showing-f10f5019