DistantNews
Support us
🇮🇱 Israel /Health & Science

AI doesn't replace good writers, it demands better judgment, researchers find

From Jerusalem Post · () English

Translated from English, summarized and contextualized by DistantNews.

At a glance

News Sources not specified Context piece
  • A study found that students using AI for writing need to exercise better judgment, challenging the idea that AI simplifies the process.
  • Researchers identified that effective AI use requires an experimental approach, human expertise, and maintaining the writer's agency.
  • AI-generated text can appear accurate but contain factual errors, necessitating students to verify information and guide the AI with context.

Students using artificial intelligence for writing may need to think more carefully, not less, according to an Iowa State University study. This challenges the assumption that AI writing tools inherently make the task easier. The research examined an "AI and Writing" course involving 38 undergraduate students from 22 different majors.

that writing with AI is an experimental process, that it requires human expertise and dialogue, and that it should strengthen, rather than replace, the writer’s own agency.

— Researchers Abram Anders and Emily Dux SpeltzThe researchers identified essential concepts for productive AI use in writing.

Researchers Abram Anders and Emily Dux Speltz analyzed student reflections on their AI writing exercises. They identified three key concepts for productive AI use: viewing AI writing as an experimental process, recognizing the necessity of human expertise and dialogue, and ensuring AI strengthens rather than replaces the writer's own agency. This perspective complicates the debate around AI in education, suggesting it is not necessarily a threat to academic integrity if students are trained to use it effectively.

AI-generated writing can often appear reliable when it is not.

— Researchers Abram Anders and Emily Dux SpeltzThe study highlighted a key problem with AI-generated text: its potential for factual inaccuracy despite fluent prose.

A central finding was that AI-generated writing can often appear reliable when it is not. This "linguistic fluency" problem means AI can produce polished prose that lacks factual accuracy. To address this, students participated in exercises like "Create a Fluent Hallucination," where they generated deliberately false but plausible AI outputs. Other assignments included prompt competitions and ethics exercises.

AI output improved when students brought their own expertise to the exchange.

— Researchers Abram Anders and Emily Dux SpeltzThe researchers observed that student success with AI writing increased when they provided context and correction.

One journalism student found ChatGPT helpful for writing leads but noted it did not adhere to journalistic structure or rules. The AI's output improved significantly only after the student provided context on journalism lead writing standards and asked for feedback on their own work. This example suggests that AI output is enhanced when students apply their own expertise, using AI as a tool that requires context and correction rather than a simple search engine.

AI still requires students to retain control over the most important parts of writing.

— Researchers Abram Anders and Emily Dux SpeltzThe study concluded that human oversight remains crucial in the AI-assisted writing process.
DistantNews Editorial

Originally published by Jerusalem Post in English. Translated, summarized, and contextualized by our editorial team with added local perspective. Read our editorial standards.

More Stories

From Our Blog