Elections are messy. As the day of any election gets closer, more stories are leaked as opponents and muck-raking journalists try to hurt the opposition’s campaign. This process can be ugly—just ask John Adams and Thomas Jefferson—but it plays an important role in informing voters. It was on display just a few weeks ago, when President Trump’s tax returns were leaked and published to call to light the fact that he had limited his federal income tax liability to just $750 in 2016 and 2017.

Similarly, a blockbuster report from the New York Post this week asserted that Joe Biden’s son, Hunter Biden, introduced business contacts to his father in an allegedly improper pay-for-influence set-up. The Post reported on “a message of appreciation that Vadym Pozharskyi, an adviser to the board of Burisma, allegedly sent Hunter Biden on April 17, 2015, about a year after Hunter joined the Burisma board at a reported salary of up to $50,000 a month.” It also reported on salacious images of Hunter Biden using drugs and engaging in sexually explicit acts.

The report has been contested, declared false by the Biden campaign, and challenged by left-leaning media outlets. The Washington Post fact-checker, for example, has cast doubt on the authenticity of the emails and the claim that Biden ever met with Pozharskyi.

Does any of this matter? Are the allegations legitimate? It should be up to voters to weigh conflicting reports and decide for themselves.

But instead, Silicon Valley tech wizards decided to throttle and ban the Post story, which they presumptively and unilaterally declared “misinformation,” across their massive platforms. Facebook used its algorithm to try and limit the story’s reach, while Twitter took it several steps further.

Twitter actually locked the New York Post’s Twitter account and even banned the president’s press secretary for sharing the story. Journalists who shared the link awoke to find their accounts suspended. The company claimed the article violated its policy on spreading “hacked content,” even though it is not proven that it stems from hacking—and this policy is rarely applied to other big stories.

Twitter even limited the ability of users to link to the story and attached a “dangerous content” warning to existing links. This unprecedented restriction on the free flow of journalism by Big Tech companies is concerning, and represents a rather transparent attempt to protect the Biden campaign. After all, data do show that employees at tech giants overwhelmingly skew Democrat in their political donations.

Yes, it’s true that Facebook and Twitter are private companies, and have the legal right to limit content on their platform. The First Amendment does not apply to them. But their decision to do so is still morally wrong and counterproductive.

The American political system depends on the free flow of information. It’s important that voters can share, debate, and contest information freely, and not have to navigate a web of censorship on the main mediums we’ve come to rely on in order to do so.

Facebook and Twitter’s efforts were also counterproductive. Their attempted crackdown totally, woefully failed.

The story about alleged corruption in the Biden family was initially making something of a splash, but wasn’t totally dominating the news. Yet once Big Tech tried to censor it, the report exploded.

It became the prime topic of discussion on cable news, and the censorship itself trended on Twitter, calling much more attention to the story than before. It prompted top lawmakers like Sens. Ted Cruz and Josh Hawley to speak out publicly and draw even more attention to the story.

Even on Facebook itself, stories about the story’s censorship were so widely shared that they dominated the website’s rankings. As Politico’s Alex Thompson notes:

So, if you wanted to keep the story about Hunter Biden quiet, well, this was just about the worst possible way to do it.

Wanna-be censors everywhere should take note. In what’s been coined the “Streisand Effect,” censoring something, especially in the age of the internet, often has the inverse effect of drawing exponentially more attention to it.

Here’s how Psychology Today explains this phenomenon:

In 2003, a picture of Barbra Streisand’s beachfront home hit the web as part of a public collection of images displaying coastal erosion. In response, in February 2003, Streisand sued the photographer for $50 million for invasion of privacy, claiming violation of a state law aimed at the telephoto lenses of paparazzi.

Ironically, the media attention surrounding the lawsuit made the photo of her house go viral (at least in 2003 terms). In the month before the lawsuit, the picture had been downloaded only six times, including twice by her lawyers—whereas the image was downloaded more than 420,000 times during the month following the lawsuit. This paradoxical result—where an attempt to silence, suppress, or stop something backfires—was dubbed the Streisand effect.

The takeaway here is clear. Sunlight, not censorship, is the best antidote for bad ideas and bad information.

So, Big Tech should rethink its misguided move toward political censorship, and embrace its original role as providing platforms for the free flow of information. Voters are smart enough to decide for themselves.

Any more major attempts by Silicon Valley to force narratives down from up above will only backfire spectacularly. If Facebook and Twitter keep it up, free-speech competitors like Parler will continue to grow. Competition and consumer pushback, not heavy-handed government regulation, are how we will preserve a free and open internet.

This piece was originally published on FEE.org under the title Big Tech Tried to Censor the NY Post’s Hunter Biden Story. They Made it Huge Instead