If someone writes about things they think will happen, but those things never materialize, they shouldn’t just get to brush it under the rug and act like they never said it. You’ve made millions of people worried over literally nothing. That should come with reputational consequences - not just for the journalist, but also for the platform that amplified their speculation.
Now obviously, there are things worth writing about even when many unknowns remain. But in those cases, acknowledge the uncertainty - lay out the improbable worst-case scenario, the more likely outcome, and the possibility that the whole issue might just fade away. Just don’t present speculation as certainty when you can’t possibly know, or if you do then own it.
Thinking about this from a technical standpoint, it would be interesting and useful if the platforms that host online articles provided some mechanisms to (1) explicitly recognize when an article is making predictions and (2) allow/remind the author or readers to follow-up and rate the accuracy of the predictions over time. This would allow all sorts of meta analysis on the accuracy of a particular author’s predictions, on particular types of predictions, on trends in positive or negative predictions, etc.