AI, bounties and culture change, how scientists are taking on errors

A simple methodological error meant that for years researchers considered drinking moderate amounts of alcohol to be healthy. Now plenty of evidence suggests that isn't the case, but errors like this still plague the scientific literature. So, how can the scientific literature become more error-free?

In the first episode of this two-part series, to determine what steps can be taken to help clean up science we explore peer review, replications, AI and even paying people to detect errors.

This episode was written and produced by Nick Petrić Howe. Dan Fox was the editor. The Swiss horn music came from DangerLaef on Freesound. All other sound effects and music were provided by Triple Scoop Music.

Correction 25 July 2025: a previous version of this podcast incorrectly stated that one in five papers are retracted, when it should have been one in 500. This has now been updated.

Subscribe to Nature Briefing, an unmissable daily round-up of science news, opinion and analysis free in your inbox every weekday.

Hosted on Acast. See acast.com/privacy for more information.