Five days after publication of Heike Vester and Marc Timme's Correspondence on the potential environmental damage by Chile's salmon farms (Nature 465, 869; 2010), the Max Planck Society in Munich posted a press release on the 'Environmental scandal in Chile' (http://go.nature.com/EnJh9B). This gave rise to media reports that should remind scientists of their responsibility to be able to back up their comments and opinions.

The two documents generated newspaper and magazine articles globally (we analysed 13 sources; available from the authors). Most of these mixed up extracts from both texts and attributed statements to Vester and Timme's “report published in the journal Nature” that were not in the Correspondence. The news articles did not mention caveats of the salmon industry worldwide: uncertain facts, disputed values, high stakes and urgent decisions.

It is uncertainties in available information that undermine evaluations of ecosystem health, rather than the potential ecological and social risks associated with the industry at large. The management of uncertainty by scientists rests on high-quality, peer-reviewed data. Bypassing that process risks compromising the credibility of science and scientists.

As a result of Nature's position as an interlocutor of the science–society interface, this case fits the concept of 'post-normal' uncertainty, in which the science is forced to play out under heavy social and political pressure — as in climate change and overexploitation of natural resources. But science still has a duty to provide sound information so that society can find ways to adapt to a changing world.