Can AI really take over journalism? Study shows why readers still prefer human-written news articles over AI-crafted pieces
Oct 25, 2024, 12:11 IST
A recent study from the Ludwig Maximilian University of Munich has shed light on a growing concern in the world of journalism: the comprehensibility of AI-generated news articles. Despite advancements in artificial intelligence, the study found that readers still prefer traditionally crafted news pieces.
The study surveyed over 3,000 UK-based online news readers, each evaluating one of 24 news articles — half written by humans and half produced by automation. The findings? Despite editorial oversight from journalists, AI-generated articles fell short on comprehensibility compared to their human-authored counterparts. Lead author Sina Thäsler-Kordonouri noted that readers found automated articles “significantly less comprehensible” overall.
One key complaint revolved around language choice in AI articles, with readers pointing out an excess of unusual or overly complex words. Another significant sticking point was how AI-handled numbers and data. For many readers, this treatment was confusing and disrupted the flow of the article, leaving them less satisfied with the automated content.
Interestingly, readers did find parity between the two types of articles in terms of narrative flow and structure. However, the challenging word choice and awkward handling of data remained a stumbling block for AI-produced articles.
Professor Neil Thurman, who headed the project, emphasised that refining the automated content process could help bridge the readability gap. This includes not only simplifying language but also finding ways to make data-driven content more intuitively understood by readers.
This research highlights the nuanced role of human editors in AI content creation. As newsrooms consider AI’s potential, balancing efficiency with clarity remains key. “Our results indicate the importance not only of maintaining human involvement in the automated production of data-driven news content but of refining it,” said Thäsler-Kordonouri.
While automation is here to stay, this study makes one thing clear: a human touch still plays a critical role in making news accessible and engaging.
Advertisement
The study surveyed over 3,000 UK-based online news readers, each evaluating one of 24 news articles — half written by humans and half produced by automation. The findings? Despite editorial oversight from journalists, AI-generated articles fell short on comprehensibility compared to their human-authored counterparts. Lead author Sina Thäsler-Kordonouri noted that readers found automated articles “significantly less comprehensible” overall.
One key complaint revolved around language choice in AI articles, with readers pointing out an excess of unusual or overly complex words. Another significant sticking point was how AI-handled numbers and data. For many readers, this treatment was confusing and disrupted the flow of the article, leaving them less satisfied with the automated content.
Interestingly, readers did find parity between the two types of articles in terms of narrative flow and structure. However, the challenging word choice and awkward handling of data remained a stumbling block for AI-produced articles.
Professor Neil Thurman, who headed the project, emphasised that refining the automated content process could help bridge the readability gap. This includes not only simplifying language but also finding ways to make data-driven content more intuitively understood by readers.
This research highlights the nuanced role of human editors in AI content creation. As newsrooms consider AI’s potential, balancing efficiency with clarity remains key. “Our results indicate the importance not only of maintaining human involvement in the automated production of data-driven news content but of refining it,” said Thäsler-Kordonouri.
Advertisement
While automation is here to stay, this study makes one thing clear: a human touch still plays a critical role in making news accessible and engaging.