Blind Tests Show People Can't Distinguish AI Writing—Until Labels Bias Them

An ACL 2025 Findings study across rewriting, summarization, and persuasive writing tasks shows people cannot reliably distinguish AI from human writing in blind tests and often prefer AI output—but preferences shift by over 30% once content is labeled as AI-generated.

2026-09-10 ~ 2026-09-10 · 2 related posts

• People prefer AI writing in blind tests—until they're told it's AI — ipeirotis · 2026-09-10
• ACL 2025 Study: Raters Can't Tell AI Text From Human, But Prefer Anything Labeled "Human" by 30% — ipeirotis · 2026-09-10