The growing need for voters to watch out for malicious AI
We knew this was going to happen. Indeed, has been happening for a while now, but this was more egregious.
A senior Trump campaign aide posted on X a video showing NBC News’ senior Capitol Hill correspondent Garrett Haake previewing the most recent Republican primary debate.
The first part of the video shows Haake doing the normal preview, but then the video cuts away to an image of Florida Governor Ron DeSantis and one hears Haake say: “This is Ron DeSantis, an establishment RINO who wears lifts in his shoes to look taller.” Then there is an image of former UN Ambassador Nikki Haley and one hears Haake say: “Nobody gives a (expletive) about Nikki Haley.”
The Trump aide who posted it, Chris LaCivita, tweeted “Now this is reporting!”
Of course, Haake never made any of those disparaging comments about the two Republican candidates chasing Donald Trump. His words – in his voice – were generated by artificial intelligence (AI), which is becoming more sophisticated by the day.
NBC’s lawyers immediately moved to have the video removed and LaCivita tweeted “to keep the NBC News lawyers off my ass…This is a PARODY.”
Some people who saw the video probably figured out pretty quickly that Haake never said those things, but everyone? And still, the damage was done.
There are tons of AI generated fakes all over the Internet as we head into the primary season. (Iowa is only a month away) One can find videos of Donald Trump or Vladimir Putin being arrested. One can even find a video of Donald Trump sitting at Ralph Kramden’s kitchen table as an irate Kramden screams at Trump “You are a blabbermouth!”
We are not going to be able to stop all of this. We know troll farms in Russia, China, Iran, North Korea and elsewhere will be busy cooking up all kinds of fake images showing the candidates saying all kinds of outrageous things (though, actually, Donald Trump does not need any help with that).
We know that the campaigns are using AI to generate first drafts of press releases, but the LaCivita video crossed a line, albeit a line that is blurry and is constantly moving.
What is required is for all the campaigns to publicly pledge that they will not engage in the practice of using AI to create fake videos, particularly those that put damaging words into the mouths of the opposing candidates and then posting them all over social media.
Of course, beyond the foreign actors, there are plenty of partisans in the U.S. who will be generating such fakes as well. There is a danger of turning the campaign for president, as well as for Congress, into an unholy mess of falsehoods that can most certainly pervert the democratic process.
Do we need little symbols on campaign videos that indicate that the videos are AI free?
Of course, this will be one more task for the news media to take on. They will have to expose these fakes, but as Mark Twain observed, “a lie can travel halfway around the world while the truth is still putting on its shoes.” Twain never dealt with social media. Those lies now can circumnavigate the globe multiple times in an instant.
But, at the end of the day, it will be largely up to the voters to sort fact from fiction. That has become increasingly harder, but even average voters have social media tools available to push back within their circles of friends and colleagues. And it would be nice if politicians who use AI to blatantly lie pay a political price.
• Keith Peterson, of Lake Barrington, served 29 years as a press and cultural officer for the United States Information Agency and Department of State. He was chief editorial writer of the Daily Herald 1984-86.