Search
A collaborative analysis report issued by the BBC alongside 22 European public media entities reveals a concerning trend: across 18 countries and 14 linguistic contexts, AI chatbots, when tasked with accessing news organization content and responding to pertinent queries, deliver approximately 45% of their answers with inaccuracies. These inaccuracies encompass factual misstatements, ambiguous source attributions, and omitted contextual details, all of which significantly undermine the credibility of the information provided. The study underscores that the processing of news by AI chatbots is frequently marred by errors stemming from the quality of training datasets and deficiencies in algorithm design. Furthermore, these systems lack genuine comprehension abilities, making them susceptible to misconstruing or even fabricating information.
-
C114 Communication Network -
Communication Home
