On March 15, 2023, a video of Ukrainian President Volodymyr Zelensky
surfaced
on Facebook, Twitter, and YouTube. In it, Zelensky announced Ukraine’s surrender to Russia and called for his people to lay down their arms. It was a turning point in world history — not because of what Zelensky said, but because he didn’t say it at all.
The video was a “deep fake” created by artificial intelligence (AI) at the direction of pro-Russia hackers. Thankfully, it wasn’t a very convincing one, and both viewers and media moderators quickly labeled it as false. But it won’t always be that easy. On the contrary, the danger of AI will only increase as time goes by. It’s a threat I’ve been
warning about
since 2018, but one that is evolving rapidly.
DEBT LIMIT TALKS SHOULD EXCLUDE NECESSARY DEFENSE SPENDING
Large language model (LLM) AI is already leaving a footprint on American civil society. ChatGPT, for example, mimics human writing so effectively that it is amplifying students’ ability to cheat in school beyond teachers’ ability to detect its traces in assignments. Meanwhile, hackers have used voice-editing programs to
portray
President Joe Biden, former President Donald Trump, and former President Barack Obama playing video games.
Neither high-school cheating nor dumb humor rises to the level of a national security threat. But
manufactured footage
of the White House announcing a military draft?
False imagery
depicting Trump’s arrest? In a country where much of our politics is driven by mass online media, deep fakes such as these, if distributed widely and marketed as true, could do tremendous damage.
Imagine what our adversaries could do with even more convincing AI. They could effectively mimic government leaders, as the pro-Russia hackers attempted to do with Zelensky, sowing confusion and chaos. They could create propaganda more quickly and distribute it more precisely. And according to the Brookings Institution’s Darrell West, they could tilt the balance in American elections by
targeting
important voting groups with “the exact message that will help them reach their final decisions.”
The tools to enact these operations are already in development, and the Chinese Communist Party — not to mention the Kremlin, Iran’s regime, cybercriminals, terrorist organizations, and more — is
planning
to put them in motion, if it hasn’t done so already. And the truth is that America is not ready to counter these threats.
Sure, policymakers and regulators are increasingly aware of the problem. The Biden administration, for instance, has announced the existence of a United States government AI content aggregator designed to identify fakes. But that’s not enough because of two distinct challenges we face.
First, there is a very limited understanding of LLM and AI generally inside government. The only experts are the people who create the actual programs — the staff of companies such as Open AI and Scale AI. The pace of development is stunning and almost impossible to keep up with. As with everything in the tech space, innovation is outrunning government.
Second, even if government could keep up, our institutions would not be up to the task of effective regulation. For years, Americans heard false claims from Democrats that Trump colluded with Russia to win the 2016 election. They saw false claims of interference again in 2020, when the media itself colluded with former intelligence officials to silence a damaging story about the Biden family. Put simply, partisan Democrats are undermining efforts to counter computer-generated lies and propaganda by using the labels of misinformation and disinformation to censor and suppress viewpoints with which they disagree.
CLICK HERE TO READ MORE FROM RESTORING AMERICA
We can and should do better. The U.S. must lead the world in adapting to the AI future and partner with technological trailblazers to make the free world more resilient to foreign influence campaigns. And policymakers must put America’s long-term interests over short-term party politics. I know that’s asking a lot of Washington, but it’s what Americans need and deserve.
Remember, we are in the midst of a great power conflict with the Chinese Communist Party. Getting AI right isn’t just about protecting ourselves from sophisticated influence operations. It’s also about determining whether the 21st century is dominated by a tyrannical regime or defined by American values. The stakes couldn’t be higher.
Marco Rubio is Florida’s senior U.S. senator.
