This just happened!😳 Read the Caption for more👇 As AI-generated text improves, it’s becoming…
This just happened!😳 Read the Caption for more👇 As AI-generated text improves, it’s becoming harder to tell the difference between content written by humans and content created by AI. A study by the University of Pittsburgh looked into whether people could reliably distinguish between poems written by well-known human poets and those generated by AI. The researchers conducted two experiments with non-expert readers and found that participants could only identify AI-generated poems with 46.6% accuracy—worse than random guessing. Interestingly, people were more likely to think the AI poems were written by humans than the actual human-written poems. The study also revealed that AI-generated poems were rated more positively in terms of rhythm and beauty. This led participants to mistakenly believe they were human-authored. One key finding was that participants used flawed reasoning to tell the difference between AI and human poetry. They often preferred the simplicity of AI-generated poems, which were easier to understand. In contrast, they sometimes misinterpreted the complexity of human poems as confusion or incoherence, mistakenly attributing it to AI.
By 5aitec
Published
From the Tech buzz AI community. Community for AI enthusiastic for leaning new algorithms
Posted in Tech innovations. Innovation at its best
3 views
0 likes