Garry Kasparov and Steve Bannon don't appear to have much in common.
One ran the chess boards, the other ran Donald Trump's campaign. One hosts the Oslo Freedom Forum, the other hosts an apocalyptic web show. One champions democracy, the other seeks to dismantle it. But for all their differences, both have a keen understanding of how propaganda works in the 21st century.
In 2016, Kasparov warned that "the point of modern propaganda," is not merely to misinform but "to exhaust your critical thinking, to annihilate truth." Bannon put it more crudely in 2018, arguing that "the real opposition is the media, and the way to deal with them is to flood the zone with shit."
It increasingly looks like they were right.
Minute after minute, hour after hour, a firehose of shit spews onto our screens and into our heads. Our information pipes are brimming with a toxic sludge of lies and conspiracy theories about everything from elections to vaccines—and it's getting harder for us to spot it. Meanwhile, as fake news accumulates, trust in real news is at an all-time low.
But new developments in artificial intelligence raise the question of whether we've only hit the tip of the iceberg.
With an estimated 100 million monthly active users in January 2023, the AI chatbot ChatGPT is now the fastest-growing app in history.
There's a lot about the tool's impact that remains unknown. Is this the death of the college essay? Or will AI enhance the learning experience, not replace it? Are knowledge workers out of a job? Or will they become more productive with AI tools?
Unfortunately, there's one domain where ChatGPT will be extremely disruptive, extremely fast: dispensing disinformation. The tool can't tell what information is true and what isn't, but it can generate boatloads of believable nonsense, making it a potent producer of political propaganda.
Even with safeguards built in to prevent its abuse, the chatbot was happy to oblige me when I asked it to "write a story about a politician named Joe Biden who grew up in a communist household in the tone of a New York Times reporter."
Here are the first few sentences it came up with: "Joseph Biden, the 46th President of the United States, grew up in a household steeped in Communist ideology and values, a fact that has had a profound impact on his political beliefs and career. Biden was born in Scranton, Pennsylvania, to Joseph Biden Sr., a staunch supporter of the Communist Party USA and an active participant in local leftist political organizations. As a child, Biden was exposed to Marxist literature."
It doesn't matter that this story is totally bogus—what matters is that it sounds believable. With trust in news organizations in the gutter, plenty of Americans would be receptive to this story if it popped up in their Facebook feeds.
As the former publisher of The Wall Street Journal told The New York Times last month, ChatGPT "is going to be the most powerful tool for spreading misinformation that has ever been on the internet."
Given the vast implications of unleashing a prolific nonsense machine ahead of a major election, it's unsurprising that ChatGPT's creator, OpenAI, waited until after the 2022 midterms to release it. But the genie is now out of the bottle, and generative AI will be everywhere by the next election.
Google is racing to release a competitor. Not only can anyone now pay to use ChatGPT in their own tools, but the tech will soon be incorporated into productivity apps like Microsoft Word. And there's more—just this week, OpenAI released a major upgrade to the model that powers ChatGPT, stating that it "can solve difficult problems with greater accuracy, thanks to its broader general knowledge and problem solving abilities."
More concerningly, malicious actors will be among the most creative and devoted adopters of these tools. Despite the fact that ChatGPT's safeguards are already easily circumvented, members of Dark Web cybercrime forums have shared tools designed to bypass them completely. Recently the model behind Facebook's ChatGPT competitor was leaked on the online forum 4chan—a favorite hangout of white supremacists, neo-Nazis, and other terminally-online alt-right communities.
The people who have been breaking our brains for years will be back with a vengeance next cycle, all but ensuring the internet will be teeming with AI-generated disinformation come November 2024.
There are more questions than answers about how to address this threat, but it is up to the forces of truth and democracy to plan for and respond to it. We cannot afford to be blindsided like we were in 2016 with the seemingly sudden rise of Russian trolls, QAnon, and fake news.
As journalist Ezra Klein put it recently, tools like ChatGPT are "going to drive the cost of bullshit to zero," making it easier than ever to flood the zone and annihilate truth. It's Steve Bannon's dream and Garry Kasparov's nightmare. And it isn't our future, but our present.
Hugh Jones works to support a mainstream, big-tent Democratic Party at The Welcome Party and writes Reality Lab, a newsletter about the intersection of politics, technology, and truth. Find him on Twitter @hughjonesiv.
The views expressed in this article are the writer's own.
Uncommon Knowledge
Newsweek is committed to challenging conventional wisdom and finding connections in the search for common ground.
Newsweek is committed to challenging conventional wisdom and finding connections in the search for common ground.
