The free online service simplifies online voice and text chat, while allowing gamers to build communities of like-minded fans. Discord lets users create specialized channels, called “servers,” dedicated to their favorite topics. Any user can create multiple servers; these can be open to all Discord users or kept private so that only friends or family members can gain entry.
By 2019, Discord was attracting 51 million unique users worldwide every month. But the company’s popularity soared during the COVID pandemic, as homebound students and workers used the network to keep in touch. By 2021, it had 100 million monthly users and Microsoft was offering to pay up to $10 billion to buy the company, an offer Discord’s founders refused.
Today, Discord claims 150 million users. That puts it behind Twitter, which has about 450 million, and Facebook, with just under three billion monthly users. Discord has raised about $1 billion in venture funding and does not release its financial data, so it’s unclear if the company turns a profit.
If Teixeira wanted to blow the cover off the deepest US secrets, Discord is a lousy way to do it. Facebook and Twitter are designed to spread news quickly. They use algorithms to show funny or shocking stories to users who otherwise might never hear about them. That’s how stories on these platforms can be viewed by millions in just a few hours.
But Discord doesn’t use algorithms to promote particular messages. Instead, its users only see what’s happening on their favorite servers. That’s how US secrets could be posted for days, unbeknownst to Discord or the government.
“Discord is actually not a great choice if you want to disseminate something quickly,” said Jason Davis, research professor at Syracuse University’s Newhouse School of Public Communications. “Usually when they do choose a platform like this, it’s because they are trying to impress what they consider their tribe, their close compatriots.”
The Teixeira crisis isn’t the first of its kind for Discord. In 2017, white supremacists used Discord servers to organize a protest in Charlottesville, Va., that led to a woman’s death. At the time, the company made little effort to control what people said on the service. Since then, it’s adopted automated content moderation tools designed to recognize and delete hateful messages.
But it’s hard to see how Discord or any other social network could be programmed to flag postings of classified information.
“I can design a system where if you use the words ‘top secret,’ it’ll delete your stuff,” said Bruce Schneier, a fellow at the Berkman Klein Center for Internet & Society at Harvard Law School, “but that would be dumb.”
Schneier said it’s relatively easy to train an AI to accurately block hate speech by feeding it thousands of hateful online messages. But it could only be trained to filter out government secrets by training it on a stack of government secrets, and that’s not going to happen.
Davis said that rather than focus on Discord, US intelligence officials should be trying to figure out how to harden their own security practices, to make it impossible for a low-ranking soldier to steal such sensitive data.
As for trying to prevent its publication after it’s been stolen, Schneier said that could only be done by abolishing freedom of speech.
“As long as human beings can talk to each other, they will share secrets,” he said. “This is not a tech story. It’s a human story.”
Hiawatha Bray can be reached at hiawatha.bray@globe.com. Follow him on Twitter @GlobeTechLab.
