“THERE’S too much sensationalism, misinformation and polarisation in the world today,” lamented Mark Zuckerberg, the boss of Facebook, recently. To improve things, the world’s largest social network will cut the amount of news in users’ feeds by a fifth and attempt to make the remainder more reliable by prioritising information from sources which users think are trustworthy.

Many publishers are complaining: they worry that their content will show up less in users’ newsfeeds, reducing clicks and advertising revenues. But the bigger problem with Facebook’s latest moves may be that they are unlikely to achieve much—at least if the flourishing of fake news on WhatsApp, the messaging app which Facebook bought in 2014 for $19bn, is any guide.

In more ways than one, WhatsApp is the opposite of Facebook. Whereas posts on Facebook can be seen by all of a user’s friends, WhatsApp’s messages are encrypted. Whereas Facebook’s newsfeeds are curated by algorithms that try to maximise the time users spend on the service, WhatsApp’s stream of messages is solely generated by users. And whereas Facebook requires a fast connection, WhatsApp is not very data-hungry.

As a result, WhatsApp has become a social network to rival Facebook in many places, particularly in poorer countries. Of the service’s more than 1.3bn monthly users, 120m live in Brazil and 200m in India (see chart). With the exceptions of America, China, Japan and South Korea, WhatsApp is among the top three most-used social apps in all big countries.

Most of the 55bn messages sent every day are harmless, but WhatsApp’s scale attracts all sorts of mischief-makers. In South Africa the service is often used to spread false allegations of civic corruption and hoax warnings of storms, fires and other natural disasters. In Brazil rumours about people travel quickly: a mob recently set upon a couple they suspected of being child traffickers based on chatter on WhatsApp (the couple escaped).

But it is in India where WhatsApp has had the most profound effect. It is now part of the country’s culture: many older people use it and drive younger ones crazy by forwarding messages indiscriminately—sometimes with tragic results. Last year, seven men in the eastern state of Jharkhand were murdered by angry villagers in two separate incidents after rumours circulated on WhatsApp warning of kidnappers in the area. In a gruesome coda to the incident, pictures and videos from the lynching also went viral.

It is unclear how exactly such misinformation spreads, not least because traffic is encrypted. “It’s not that we have chosen not to look at it. It is impossible,” says Filippo Menczer of Indiana University’s Observatory on Social Media, which tracks the spread of fake news on Twitter and other online services. Misinformation on WhatsApp is identified only when it jumps onto another social-media platform or, as in India, leads to tragic consequences.

Some patterns are becoming clear, however. Misinformation often spreads via group chats, which people join voluntarily and whose members—family, colleagues, friends, neighbours—they trust. That makes rumours more believable. Misinformation does not always come in the form of links, but often as forwarded texts and videos, which look the same as personal messages, lending them a further veneer of legitimacy. And since users often receive the same message in multiple groups, constant repetition makes them more believable yet.

Predictably, propagandists have employed WhatsApp as a potent tool. In “Dreamers”, a book about young Indians, Snigdha Poonam, a journalist, describes visiting a political party’s “social media war room” in 2014. Workers spent their days “packaging as many insults as possible into one WhatsApp message”, which would then be sent out to party members to be propagated within their own networks. Similar tactics are increasingly visible elsewhere. Last month’s conference in South Africa of the African National Congress, at which delegates elected a new party leader, saw a flood of messages claiming victory for and conspiracy by both factions. With elections due in Brazil and Mexico this year, and in India next year, expect more such shenanigans.

Governments and WhatsApp itself are keenly aware of the problem. In India authorities now regularly block WhatsApp to stop the spread of rumours, for instance of salt shortages. Regulators in Kenya, Malaysia and South Africa have mooted the idea of holding moderators of group chats liable for false information in their groups. WhatsApp is working on changing the appearance of forwarded messages in the hope that visual cues will help users tell the difference between messages from friends and those of unknown provenance. But ultimately it will be down to users to be more responsible and not blindly forward messages they receive.

It is as yet unclear whether fake news on Facebook will be less of a problem after it changes its algorithms. The experience of WhatsApp suggests, however, that the concerns will persist. “Even with all these countermeasures, the battle will never end,” Samidh Chakrabarti, a Facebook executive admitted on January 22nd. “Misinformation campaigns are not amateur operations. They are professionalised and constantly try to game the system.”