Quartz, the news brand, is using chatbots and artificial intelligence as it seeks to personalise stories for its audience on a truly individual basis.
Emily Withrow, director of Quartz’s Bot Studio, discussed this topic during a panel session held at SXSW 2019 in Austin, Texas.
“Our challenge is to write stories that you can talk to on a number of different platforms,” she said. (For more, read WARC’s in-depth report: How Quartz is using artificial intelligence to truly personalize the news.)
“You can talk back to the stories; you can type things; you can choose options; and the story really works and adapts to you.”
A case in point involves a chatbot that is being launched for Samsung phones on AT&T’s network, with ten million devices scheduled for inclusion in the first-phase roll out.
This tool will aim to recognise whether a reader has the time to read an in-depth story or simply wants to scan the headlines, as well as learning their preferences over time.
Prior attempts at customisation in the digital-news business, Withrow argued, typically asked individuals to identify the subjects of greatest personal interest, and then offered bespoke homepages or email newsletters.
Now, however, the possibilities run considerably deeper. “One thing that we’re doing is experimenting with this on the story level,” she asserted.
“We’ve done a lot of experimentation on the site level. But when you look at the smaller components, you’re really looking at stories.”
Alongside adapting a story depending on an individual’s prior level of knowledge – “We can fill in background if you need it – or not, if you don’t,” Withrow said – this approach may be able to help in tackling political bias.
“One of the problems that I’ve been thinking a lot about is the problem of people not engaging with information that is counter to whatever they already believe,” said Withrow.
“So it could mean – and we’ve been running experiments on this – could we rearrange some of the paragraphs in these stories to show your side first and then show you the counterpoint, versus showing everyone the counterpoint first.
“What we’ve found is that, yes, people do actually get further into the story [and] engage with … a different side of the story they don’t necessarily support.”
Emily Withrow, director of Quartz’s Bot Studio, discussed this topic during a panel session held at SXSW 2019 in Austin, Texas.
“Our challenge is to write stories that you can talk to on a number of different platforms,” she said. (For more, read WARC’s in-depth report: How Quartz is using artificial intelligence to truly personalize the news.)
“You can talk back to the stories; you can type things; you can choose options; and the story really works and adapts to you.”
A case in point involves a chatbot that is being launched for Samsung phones on AT&T’s network, with ten million devices scheduled for inclusion in the first-phase roll out.
This tool will aim to recognise whether a reader has the time to read an in-depth story or simply wants to scan the headlines, as well as learning their preferences over time.
Prior attempts at customisation in the digital-news business, Withrow argued, typically asked individuals to identify the subjects of greatest personal interest, and then offered bespoke homepages or email newsletters.
Now, however, the possibilities run considerably deeper. “One thing that we’re doing is experimenting with this on the story level,” she asserted.
“We’ve done a lot of experimentation on the site level. But when you look at the smaller components, you’re really looking at stories.”
Alongside adapting a story depending on an individual’s prior level of knowledge – “We can fill in background if you need it – or not, if you don’t,” Withrow said – this approach may be able to help in tackling political bias.
“One of the problems that I’ve been thinking a lot about is the problem of people not engaging with information that is counter to whatever they already believe,” said Withrow.
“So it could mean – and we’ve been running experiments on this – could we rearrange some of the paragraphs in these stories to show your side first and then show you the counterpoint, versus showing everyone the counterpoint first.
“What we’ve found is that, yes, people do actually get further into the story [and] engage with … a different side of the story they don’t necessarily support.”
Sourced from WARC
