A study by the Tow Center for Digital Journalism revealed that AI-powered search engines are failing in properly citing news sources, crucial for maintaining public trust in journalism. Researchers tested eight AI models using 10 excerpts each from 20 publishers, manually evaluating 1,600 results. The findings highlighted a serious concern: these tools not only fabricated citations but also misidentified articles and publishers, thus disrupting the flow of traffic to original journalism. This failure raises significant implications for both the media landscape and audience trust.
Across multiple models, the study found a concerning trend: AI search engines demonstrated a tendency to fabricate citations and often misattributed original articles, ultimately undermining trust in news sources.
The research investigators emphasized that while AI search tools offer remarkable possibilities for information retrieval, they significantly fail at fulfilling the critical requirements of accurately quoting and citing news publishers.
Collection
[
|
...
]