The AI election
Political campaigns and groups are already using AI to generate ads, emails, and images. When will Washington start paying attention?
BREAKING: Hunter Biden, the 53-year-old son of President Joe Biden, has been charged with illegally possessing a handgun and failing to pay taxes for two years. The younger Biden has agreed to plead guilty to the two misdemeanor tax counts and to enter a probation agreement for the gun charge. If the plea deal is approved by a federal judge, it will likely keep him out of jail.
This news is breaking just as I’m sending out the newsletter. I will have more on this tomorrow; for now, here’s your regularly scheduled programming:
Good morning! It’s Tuesday, June 20, 2023. The 2024 elections are 504 days away. If this newsletter was forwarded to you, subscribe here. If you want to contribute to support my work, donate here.
Earlier this month, Ron DeSantis’ campaign posted a video on Twitter upbraiding Donald Trump for not firing Dr. Anthony Fauci during his presidency.
At one point, the video features a collage of images that purport to show Trump hugging Fauci, a figure despised by many in the Republican base, and even kissing him on the cheek.
The only problem? Half of them were fake.
I’ve included a screengrab from the video below; see for yourself if you can tell which images aren’t real.
If you guessed the top-left, bottom-middle, and bottom-right photos, you got it right: those aren’t real images, they were created by artificial intelligence.
The DeSantis campaign included no message disclosing that they had mixed in fake images of their opponent along with real ones. As of this writing, the video (originally uploaded on June 5) remains posted on their Twitter account.
This is just one example of how campaigns are already utilizing AI in the 2024 election.
Get ready for a whole lot more, as AI threatens to disrupt American politics just as it has already upended everything from the student essay to coding.
The potential dangers posed by AI in a political context are obvious. Not only could a campaign or outside group use DALL-E or Midjourney to steal an opponent’s likeness and create fake images of them; it’s now easier than ever to use a steal an opponent’s voice and create fake videos or audio of them.
Imagine a political ad that includes a “message from Joe Biden” or an “announcement from Donald Trump,” but the video is completely AI-generated. Or a robocall targeted at certain voters that purports to air an offensive message from a rival candidate, except the audio isn’t genuine.
Such AI-generated video or audio are known as “deepfakes,” and as you may have seen in examples online, they can be pretty convincing:
While warnings about “deepfakes” have bounced around the political sphere for years, with the rise of new AI tools and platforms, their use is no longer theoretical.
In addition to the DeSantis video, the Republican National Committee released a completely AI-generated digital ad in April, after President Biden announced his re-election bid. The ad is intended to show what a second Biden term would look like, flashing a series of dystopian images — Taiwan being attacked, a surge of migrants at the border, a violent gang — that were all created by AI.
A small disclosure hovers in the top-left corner (“Built entirely with AI imagery), but it’s easy to miss. (Personally, I only spotted it on my third rewatch.)
On the other side of the political aisle, the Democratic National Committee has already started experimenting with using AI to draft its fundraising emails. (Just what we needed: more political emails asking for money!)
And AI-generated misinformation has been making the rounds lately: last month, stock markets briefly dipped as an image spread on Twitter that appeared to show an explosion near the Pentagon. Accounts with hundreds of thousands of followers — including RT, Russia’s state media channel — tweeted the photo, before it was eventually revealed to be a fake.
AI ads and photos have been able to spread so easily partially because the space remains so unregulated by policymakers. As usual, Washington runs at least five years behind Silicon Valley — if not more.
But the White House and lawmakers in both parties are beginning to take notice of AI, including with several notable pushes this week:
Today, President Biden (who has reportedly used ChatGPT) will meet with AI experts and researchers in San Francisco to discuss the opportunities presented by the technology for education, medicine, and other sectors — as well as “the risks posed by AI if it isn’t properly regulated,” according to the White House. Attendees will include Khan Academy founder Sal Khan, Center for Human Technology co-founder Tristan Harris, and Stanford’s Human-Centered AI Institute co-director Fei-Fei Li.
On Wednesday, Senate Majority Leader Chuck Schumer (D-NY) is set to deliver remarks at a think tank in D.C. to unveil his “SAFE Innovation framework” for artificial intelligence. According to Schumer, his plan will present an outline for “comprehensive” and “bipartisan” AI legislation he hopes to soon pass through the Senate. (Schumer and two Republican colleagues also recently hosted the first-ever all-senators briefing on AI.)
Later this week, Reps. Ted Lieu (D-CA) and Ken Buck (R-CO) are set to introduce a resolution to create a bipartisan commission that will provide recommendations on regulating AI. Lieu penned a New York Times op-ed in January teasing the resolution, titled: “I’m a Congressman Who Codes. AI Freaks Me Out.”
This week’s push presents the possibility that Washington could be preparing to assume a larger role in AI regulation, a province that has so far been left mostly to the states.
In 2019, Texas became the first state to criminalize deepfakes, implementing a law that prohibited campaigns from publishing deepfake videos “with intent to injure a candidate or influence the result of an election” in the last 30 days before an election. Violating the law can lead to a year-long sentence in county jail and a fine up to $4,000.
Several other states — from Republican-led Wyoming to Democratic-led New York — have also enacted laws cracking down on deepfakes, as well as other AI regulation measures.
A flurry of bills have been introduced in Washington in response, including the REAL Political Advertisements Act by Sen. Amy Klobuchar (D-MN), which would require campaigns to disclose their use of AI-generated content in advertising.
So far, none of the proposed federal measures have been enacted.
More news you should know
Judge Aileen Cannon has set a preliminary start date of August 14 for the federal criminal trial of Donald Trump. The date would set an aggressive timeline for the highly anticipated proceedings, but don’t expect it to stick: Trump’s lawyers are expected to file several motions to attempt to delay the trial’s opening. Read Cannon’s order here
Last night, in an interview with Fox News anchor Bret Baier, Trump seemed to admit that he kept classified documents even after a grand jury subpoenaed him for them. “I was very busy,” he offered as an excuse; he also said he refused to comply because the “boxes were interspersed with all sorts of things: golf shirts, clothing, pants, shoes” in addition to top-secret military documents.
Trump also said he was not worried at all about his culpability in the case, based on the Presidential Records Act and the “Clinton socks case.” I explained last week why neither of those defenses will help Trump much in court.
The FBI resisted opening an investigation into Trump’s role in the January 6th riot for over a year, according to the Washington Post. Although fearful of appearing partisan, the bureau eventually reversed itself and began examining the ex-president and his allies, partially spurred to action by the work of the House January 6th committee.
China and Cuba are “negotiating to establish a new joint military training facility” on Cuba’s northern coast, according to the Wall Street Journal. The facility is separate from a secret Chinese spy base that Cuba has reportedly also agreed to host; both developments have sparked fears in Washington, amid broader tensions between the U.S. and China. The news comes one day after both sides said that “progress” was a made in a meeting between Secretary of State Antony Blinken and Chinese leader Xi Jinping in Beijing.
A few more headlines...
Supreme Court considers recoil from landmark gun rights ruling (NBC)
2024’s hidden prize: The upper hand in tax ‘Armageddon’ (Politico)
U.S. Becomes Transgender-Care Outlier as More in Europe Urge Caution (WSJ)
Today’s political planner
All times Eastern.
President Biden is in California. He will deliver remarks at 4 p.m. in San Francisco on his “commitment to seizing the opportunities and managing the risks of artificial intelligence.” Later tonight, he will headline campaign fundraisers in San Francisco and nearby Kentfield.
Vice President Harris is in Dallas. She will tape a roundtable conversation on abortion for MSNBC’s “The ReidOut” at 4:15 p.m. and headline a campaign fundraiser at 6:30 p.m. before returning to Washington.
First Lady Biden will host a roundtable conversation at 3:45 p.m. with women who have been denied medical care since Roe v. Wade was overturned. Both the VP’s and the first lady’s events are part of a series of White House events to mark the one-year anniversary of Roe’s reversal, which is Saturday.
The Senate is scheduled to vote at around 5:30 p.m. to confirm Julie Rikelman as a U.S. Circuit Judge for the First Circuit.
The House is scheduled to vote on a Senate-passed bill authorizing the appropriation of $3.7 billion to construct or renovate seven major Veterans Affairs medical facilities this fiscal year. The chamber will also vote on a bill establishing “Boots to Business,” an entrepreneurship training program for veterans, and a bill renaming a VA clinic in Michigan.
The Supreme Court will release orders — indicating the cases it does and does not plan to take next term — at 9:30 a.m.
Republican presidential candidate Tim Scott will participate in a Fox News town hall with Sean Hannity. Fox’s second town hall of the 2024 cycle, the event will take place in South Carolina and air at 9 p.m.
Before I go...
Here’s something interesting: Are people in some professions more likely to marry people in certain other professions?
It turns out they are, according to the Washington Post. The most common professions to marry each other are firefighters and registered nurses: almost 10% of firefighters marry nurses, per census data.
The most common occupation to intermarry are doctors: 18.5% of married physicians and surgeons are wedded to other physicians and surgeons. Close behind are college professors and restaurant, bar, and hotel managers.
Read more via the Washington Post.
Thanks for reading.
I get up each morning to write Wake Up To Politics because I’m committed to offering an independent and reliable news source that helps you navigate our political system and understand what’s going on in government.
The newsletter is completely free and ad-free — but if you appreciate the work that goes into it, here’s how you can help:
Donate to support my work or set up a recurring donation (akin to a regular subscription to another news outlet).
Buy some WUTP merchandise to show off your support (and score a cool mug or hoodie in the process!)
Tell your family, friends, and colleagues to sign up at wakeuptopolitics.com. Every forward helps!
If you have any questions or feedback, feel free to email me: my inbox is always open.
Thanks so much for waking up to politics! Have a great day.
— Gabe