Indonesia votes in three weeks, and the effect is visible on every phone. The WhatsApp groups we belong to, including the dive crew, suppliers and neighbours, now carry a constant stream of screenshots, voice notes and "forwarded" videos. Some of this content is accurate; much of it is not. Most of it arrives with no source, no date and many exclamation marks.
In January, WhatsApp limited forwarding to five chats at a time worldwide. Around the same time, YouTube announced that it would reduce recommendations of miracle-cure and flat-earth videos. Two of the largest platforms acknowledged the same issue in the same month: their systems are very effective at spreading content and much less effective at determining whether it is true.
In this region, WhatsApp functions as the main telephone system. Every business relies on it, and so does every rumour.
Are social media feeds neutral?
Very few people now see posts in the order they were published. Facebook, Instagram, YouTube and Twitter all rank content according to what they predict will prompt a reaction: clicks, comments, shares and viewing time. In principle, this is reasonable. In practice, content that provokes anger or fear generates more of all these reactions.
The feed therefore learns to show more of it. No one at these companies needed to decide that outrage was good for business; the metrics produced that result, one small test after another.
Users also play a role that is often overlooked: they train the system. Every pause on an alarming headline counts as a signal. The algorithm does not follow a master plan; it reflects user behaviour while amplifying certain tendencies.
This conclusion is uncomfortable, but it is also useful, because part of the solution lies in individual habits rather than in waiting for the technology industry to correct itself.
How can you check what you read and share?
The following routine is deliberately simple. It can be completed in the time it takes to press share, and it requires no special tools.
Who published it?
Before reading the content, check who published it. Is the name familiar? Is there an about page, an address, real people? Much unreliable content fails at this step. A site with a news-sounding name, no author and a domain registered last month is not a regional newspaper, even if the headline appears identical in the feed.
In group chats, the problem is greater, because there is often no source at all, only a screenshot of a screenshot. We apply a simple rule: no source, no share.
When was it published?
Outdated content misleads more people than outright fabrications: a genuine story from three years ago circulated as if it happened this morning, an old flood photo, a road closure resolved long ago, or a price increase that was later reversed. Feeds remove context, and the date is usually the first element lost. Careful and well-informed people are regularly misled this way, and we are no exception.
Who benefits?
This question is practical, not conspiratorial: who gains if this is believed and shared? Sometimes it is a company selling a product. Sometimes it is a page cultivating outrage to gain followers. Sometimes no one benefits and it is an honest mistake. Asking the question takes about two seconds, and two seconds is usually enough.
For a more structured approach, the News Literacy Project offers a free platform called Checkology. It is designed for students, but its lessons on distinguishing news from opinion and identifying altered photos are valuable for anyone, and many adults would benefit from them as much as their children.
Why is misinformation a small business issue?
Most current coverage of this subject focuses on elections, which is understandable. It is election season here, Canada has a federal election in October, and the concerns are real.
The business implications, however, deserve equal attention.
A business page is now a publisher, whether or not the owner intended it to be. When a shop reshares an alarming post about a local "scam" that proves to be false, its name is attached to it. When a tourism operator shares an incorrect figure, guests remember who said it. On a small island, the entire community knows by lunchtime. Trust takes years to build and can be lost in an afternoon.
The reverse is also true. A page known for its care, one that shares information that proves accurate, builds a reputation that cannot be bought. People begin consulting the page to check whether something is true, and in a small community, that has considerable value.
If you manage the page, or pay someone to manage it, we recommend agreeing on a few basic rules for what gets shared. A measure of social media management discipline is highly effective in this area. What a business shares says as much about it as what it publishes.
What risks lie ahead?
Photos and video have traditionally served as decisive evidence: if there was an image, the event probably happened. This assumption is becoming less reliable.
Last year, BuzzFeed and Jordan Peele released a fake video of Obama saying things he never said, as a warning, and it was convincing enough to make the point. The software behind these "deepfakes" is becoming easier to use, and last month OpenAI announced that it was withholding a text generator, GPT-2, because it writes convincing fake news. Taken together, these developments weaken the "does it look real" test year after year.
We expect the three basic questions to become more important: source, date, and who benefits. They do not depend on the content appearing fake, and they do not depend on any platform correcting itself.
No one is immune. Feeds are designed to capture attention when people are tired and scrolling late at night, at eleven in the evening. The most effective protection is to establish the habit in advance rather than attempting to apply it in the moment.