"Chatbots are coming to news - strike that; they're here already - and we're going to have to find a way to work in that world, despite all their well-documented limitations. This isn't an argument that chatbots are a good source of news, but it's a prediction that more and more people will turn to them because they provide something legacy newsrooms don't: personalized information that's more useful for each user."
"A recent Reuters Institute study noted that about 7% of adults already use chatbots as a source of news - but that number doubles to 15% for the under-25 crowd. How soon before it hits Facebook levels (36%) or more - and what will that mean for news, journalism, society, democracy, and more? Large language models are language models, not fact models, which partly explains why they spew hallucinations with depressing regularity."
Audiences are shifting toward chatbots for news because chatbots deliver personalized, locally relevant, and user-specific information that legacy newsrooms often do not provide. Chatbot usage is already measurable among adults and is notably higher among under-25s, suggesting potential for much wider adoption. Users often want concrete, personal impacts, such as how a federal tax proposal affects a small business in Ohio or how a school board vote affects a child's education. Large language models generate fluent language but are not inherently factual and can hallucinate. Combining LLMs with verified sources through retrieval-augmented generation can produce more precise, accurate answers and is driving structural change in news production and delivery.
#chatbots #large-language-models #news-personalization #retrieval-augmented-generation #journalism-adaptation
Read at Nieman Lab
Unable to calculate read time
Collection
[
|
...
]