It only took one click.
And then, for 11 startling minutes—or blissful ones, depending on your politics—the constant drumbeat that is the @realdonaldtrump Twitter handle was muted, taken offline Thursday evening by a Twitter customer-service worker on his or her last day.
The President, for one, seems to have taken the bold move as a compliment.
But for Twitter, the worker’s final act couldn’t come at a worse time. It exposed yet another gaping security flaw for the social platforms that now find themselves at the center of international politics. Twitter has become the leader of the free world’s preferred communication channel, and yet, customer-service agents within the company have the power to shut him up. Having such a massive blind spot exposed would be wounding for Twitter no matter when it was revealed, but the timing of this epic exit was particularly apt. This week, three congressional committees grilled Twitter, Facebook, and Google about the role their platforms played in enabling Russian propaganda to spread during the 2016 election. Lawmakers reprimanded the tech titans for their lack of foresight and urged them to think through the tremendous and unprecedented power they now hold.
Thursday’s silencing of @realdonaldtrump encapsulates the central tension of this week’s hearings, a tension that applies to all three companies: While human oversight is often held up as the solution to algorithms run amok, giving people the power to judge what can be said and who can say it can be just as fraught.
Last year, for example, critics assailed Facebook after reports that the human moderators of its Trending Topics section, which highlights trending news, were biased against conservative news outlets. Facebook got rid of the human moderators, but that posed its own problems, allowing fake news stories to automatically populate that section.
“If you want the companies to be able to have people who look at accounts and content and take action on them accordingly, that means the companies need a whole lot of people, who have the power to click a button and delete accounts,” says Adam Sharp, Twitter’s former head of government, news, and elections. “Just like policing in the real world, where you have to have people who have the authority to slap handcuffs on you, you also have people who abuse that power. That tradeoff was illustrated better by this than anything that was said in the hearings.”
In this case, it’s technology that’s needed to check the impulses of us mere mortals. Twitter says it has updated its policies to prevent this kind of abuse in the future; the company did not respond to WIRED’s request for comment.
The employees with the power to shut down accounts at Twitter are either members of the Trust and Safety team or members of the Customer Support team. Trust and Safety are the people who are alerted if you’re being harassed or doxxed. Customer Support is the team that helps if you were locked out of your account. Some are full-time employees, and others are contractors, all divided between Twitter’s San Francisco and Dublin headquarters. (The New York Times is reporting that the person who deleted Trump’s account was a contractor).
These teams can remove tweets that violate Twitter’s terms of service and also have the ability to reset an email address associated with an account. That means they could, in theory, delete a message from the president they don’t like or create a new address to change the account’s password and tweet from it. Resetting a verified account would, however, trigger an email to the President’s original email address notifying him of the change and set off a flag for review at Twitter.
No such security protocol exists for deactivating accounts. One former Twitter employee told Buzzfeed that deactivating an account only takes one click, if you have the access, and hundreds of people do.
While the suspension heard ’round the world may have come at a terrible time for tech, it could have been worse in terms of its global impact. Take, for instance, the time President Trump proposed a ban on transgender people serving in the military on Twitter. The commander-in-chief left the world in suspense with an initial tweet that read:
“If that had been the moment that the account disappeared, it would have very quickly started causing significant concern among allies,” Sharp says. As it was, the president waited nine minutes before he completed his thought, sparking panic at the Pentagon.
No doubt, the tech titans’ problems stem largely from their leaders’ failure—or refusal—to anticipate the worst ways their creations will be used. But, crucially, they’re also the result of the new ways that governments, their leaders, and President Trump in particular, are using these platforms to announce policy, denounce critics, and even threaten war. Neither the companies, nor the US government, have caught up to that shift.
Today, the President of the United States has the same little blue check mark next to his Twitter handle as Snooki, Bill Nye, and I do. In terms of power, and well, a lot of other measures, the four of us have little in common. It’s not hard to imagine a world in which Twitter categorized heads of state, or even elected officials, differently than other users, shrouding their accounts in a different level of security. Of course, Sharp notes, tech companies are always reluctant to make such judgment calls.
“As soon as you draw one line, you’re constantly in a debate about why you didn’t draw a line over there,” he says, calling this another arena where regulatory oversight is lagging in the US. There are no clear rules requiring social content to be preserved by the National Archives, for example. It’s up to individual agencies to determine whether a Tweet or Facebook post is, in fact, part of the federal record. That leaves documenting history up to the whims of social media platforms. Case in point: In September, President Trump retweeted an account that Twitter has now identified as a Russian bot. But the Russian bot tweet the President responded to is no longer visible online, because Twitter has suspended it and more than 2,700 accounts linked to it.
Democratic Congressman Mike Quigley introduced the COVFEFE Act in June, named after President Trump’s infamous, and now deleted, middle of the night Twitter typo. It would expand the definition of presidential records to include social media.
Sharp says both tech companies and the government would be wise to review their standards regarding protecting and preserving the most sensitive accounts. “The president is not just a normal person,” Sharp says.
You can say that again.