And finally tomorrow at 15:30: a joint paper with @amyrossarguedas.bsky.social, @richardfletcher.bsky.social and @rasmuskleis.bsky.social on audiences and their views on AI-generated misinformation 🤖 ⚠️

The editorial team at @reutersinstitute.bsky.social has put together a website of our talks, featuring @waqasejaz.bsky.social, @richardfletcher.bsky.social, @amyrossarguedas.bsky.social, @rasmuskleis.bsky.social, & a collab with @c3ds.bsky.social

reutersinstitute.politics.ox.ac.uk/news/here-ar...

You can find the complete analysis chapter here: reutersinstitute.politics.ox.ac.uk/digital-news...

Younger people, who tend to be more comfortable with AI in general, also showed greater interest in the use of AI for personalising *formats* (e.g. adapting articles to different reading levels), as well as chatbots.

More broadly, respondents tended to express more enthusiasm across the board in countries where comfort with the use of AI in journalism is higher, such as India and Thailand, whereas we see much lower interest in countries with low AI comfort, such as the UK.

We found some differences by markets: e.g. translation ranks higher in linguistically unique European countries (e.g. Finland, Hungary); interest in adapting language for dif reading levels ranks higher in countries with lower literacy rates or reading proficiency (most popular option in India).

We found relatively low interest – below 30% – for any single option (which may be shaped by low familiarity with these kinds of tools). Interest was highest in options for making news use more efficient (summaries and translations) and relevant (customised homepages and recommendations or alerts).

Then we looked at audience interest in 8 different types of AI-driven news personalisation, including options for personalised *selection* (e.g. customised news homepages), but also options for personalised *formats* (e.g. summaries, translations, text-to-audio) and AI chatbots.

In open comments, those who said they were comfortable with personalised news selection explained they appreciated things like greater relevance, efficiency, and diversity; those who were uncomfortable questioned the quality of recommendations or worried about biases, missing out, surveillance, etc.

We found that close to half of respondents are comfortable with personalised news selection, but comfort was lower than personalisation for weather, music, or online TV (and higher than for social media/video feeds).

First, we looked at audience comfort with personalised *selection* across different kinds of websites and apps to understand how comfort with algorithmic recommendation in news compares to other domains.

If you haven't seen them already, you can find details on all of our launch events here: bsky.app/profile/reut...