🧵 New WP 🧵

Should AI-generated content be labeled as such on social media?

We show that these labels would have detrimental consequences on harmless AI content and would be less effective than false labels in preventing the spread of false news.

osf.io/preprints/psya…
We conducted a pre-registered online experiment with 1,975 US respondents.

In a between-subjects design, participants rated their sharing intentions and perceived accuracy of 16 headlines. Participants were randomly assigned to one of the following conditions.

🚨 New paper in @MisinfoReview 🚨

We surveyed 150 experts on misinformation and identified areas of expert consensus regarding definitions of misinformation, its determinants, solutions, and the future of the field.



🧵 https://t.co/Jb55XpcJ4ymisinforeview.hks.harvard.edu/article/a-surv…
1) Definitions

Experts defined misinformation as “False and misleading information” or “False and misleading information spread unintentionally”.

Qualitative experts were more likely to include intentionality in the definition than quantitative researchers.

🚨 New preprint 🚨 With @hugoreasoning & Emma we identified a factor that could explain why people share news they suspect to be inaccurate (such as fake news): the ‘interestingness-if-true’ of a piece of news.

Check it out: psyarxiv.com/tdfh5/

1/10 Think about it: how interesting would it be if COVID-19 was effectively a bioweapon released by China? Or if the moon was populated by “bison, goats, unicorns, bipedal tail-less beavers and bat-like winged humanoids”?

2/10

Here is a thread with some key facts about fake news that we presented with @berriche_manon at the Festival of Science:

1/6 Fake news are consumed by a minority of internet users (whether it’s on Twitter or Facebook) and are shared by an even smaller minority (0.1%).

Sources:
ncbi.nlm.nih.gov/pubmed/30679368
advances.sciencemag.org/content/5/1/ea…
papers.ssrn.com/sol3/papers.cf…
science.sciencemag.org/content/363/64…
2/6