Why AI Didn't Impact the 2024 Elections

0:00Today on the AI Daily Brief, why AI didn't really impact the 2024 elections. The AI Daily Brief is a daily podcast and video about the most important news and discussions in AI. To join the conversation, follow the Discord link in our show notes.

0:17Welcome back to the AI Daily Brief. Today is, of course, election day in the United States. That means a lot of things to a lot of people, many of which we will not be getting into on this show. However, there are some interesting things to explore when it comes to artificial intelligence. Part of that is, of course, policy looking forward. AI is likely to be one of the most consequential issues faced by the next president, whether it's Kamala Harris or Donald Trump. The Biden White House has made it clear that AI is a national security priority. All indicators suggest that the U.S. is getting more and more convinced that competitiveness and global leadership in AI is a major strategic imperative.

0:53AI is already finding its way into policy vis-a-vis China and trade restrictions. And what's more, on top of all of this, over the next four years, the capabilities of AI are likely to continue to grow, making it just ever more significant to the structure of society and the economy. Yet today what I want to explore is the election itself. God willing, this will be the last day of the election, although I'm not overly optimistic about that. And it's useful to maybe go back and remember just how terrified people were of this exact moment because of artificial intelligence. What's on your screen right now, and apologies if you're just listening, but I'll describe it, is Google search results for AI elections from November 2022, which is of course when ChatGPT launched, to about six months later in June of 2023.

1:38There are infinite thought pieces about this. The Brennan Center for Justice, how AI puts elections at risk and the needed safeguards. Brookings, how AI will transform the 2024 elections. On PBS.org, AI-generated disinformation poses threat of misleading. The conversation, how AI could take over elections and undermine democracy. Axios, how AI is already changing the 2024 election. AP, AI presents political peril for 2024 with threat to mislead. The New York Times, AI's use in elections sets off a scramble for guardrails. Scientific American, how AI could take over elections, blah, blah, blah, blah, blah.

2:13You get the point. The point is that there was a lot of concern that AI was going to be very, very bad for this election cycle and that it would be the beginning of a persistent challenge. And by the way, this wasn't just from anti-tech folks. Ex-Google CEO and prolific AI investor Eric Schmidt said that AI would make the 2024 elections a mess. For him, it was all about the social media platforms not taking responsibility, saying that they were not protecting us from false generative AI. Now, of course, the fear was that AI was going to create widespread access to incredibly powerful, high-quality generative audio, image, and video, making it incredibly cost-effective and easy to run large-scale disinformation campaigns.

2:52This fear was really ratcheted up when generative video started getting good at the beginning of this year. In February, for example, around the first previews of OpenAI's Sora, experts became terrified of the implications. Oren Etzioni, the founder of non-profit truemedia.org, said, Generative AI tools are evolving so rapidly, and we have social networks, which leads to an Achilles heel in our democracy, and it couldn't have happened at a worse time. As we're trying to sort this out, we're coming up against one of the most consequential elections in history. The threat was perceived to be so real that the Department of Homeland Security put out a bulletin in May.

3:24They warned that, quote, A variety of threat actors will likely attempt to use generative artificial intelligence augmented media to influence and sow discord during the 2024 U.S. election cycle. And AI tools could potentially be used to boost efforts to disrupt the elections. As the 2024 election cycle progresses, generative AI tools likely provide both domestic and foreign threat actors with enhanced opportunities for interference by aggravating emergent events, disrupting election processes, or attacking election infrastructure. And then, of course, there was historian and sapiens author Yuval Noah Harari.

3:54He had some of the most bombastic and dramatic warnings. Addressing the UN's AI for Good Global Summit in Geneva, he said, Now it is possible for the first time in history to create fake people, billions of fake people. If this is allowed to happen, it will do to society what fake money threatened to do to the financial system. If you can't know who is a real human, trust will collapse. Maybe relationships we'll be able to manage somehow, but not democracy. Now, presumably he's talking about crypto there. And at the risk of blending the streams of these podcasts, I'm not totally sure. The analogy holds a ton of water given that crypto is bigger than it's ever been.

4:26And nothing in the economy seems to be breaking because of it, but that's neither here nor there. Anyways, the point is that Harari is very concerned and isn't really sure about whether democracy and AI can coexist. Today's episode is brought to you by Vanta. Whether you're starting or scaling your company's security program, demonstrating top-notch security practices and establishing trust is more important than ever. Vanta automates compliance for ISO 27001, SOC 2, GDPR, and leading AI frameworks like ISO 42001 and NIST AI risk management framework, saving you time and money while helping you build customer trust.

4:59Plus, you can streamline security reviews by automating questionnaires and demonstrating your security posture with a customer-facing trust center all powered by Vanta AI. Over 8 ,000 global companies like LangChain, Lila AI, and Factory AI use Vanta to demonstrate AI trust and prove security in real time. Learn more at vanta.com slash nlw. That's vanta.com slash nlw. Today's episode is brought to you by Plum. Want to use AI to automate your work but don't know where to start? Plum lets you create AI workflows by simply describing what you want. No coding or API keys required. Imagine typing out, AI, analyze my Zoom meetings and send me your insights in Notion and watching it come to life before your eyes.

5:37Whether you're an operations leader, marketer, or even a non-technical founder, Plum gives you the power of AI without the technical hassle. Get instant access to top models like GPT-40, Cloud Sonnet 3.5, Assembly AI, and many more. Don't let technology hold you back. Check out Use Plum, that's Plum with a B, for early access to the future of workflow automation. Today's episode is brought to you by Superintelligent. Every single business workflow and function is being remade and reimagined with artificial intelligence. There is a huge challenge, however, of going from the potential of AI to actually capturing that value.

6:11And that gap is what Superintelligent is dedicated to filling. Superintelligent accelerates AI adoption and engagement to help teams actually use AI to increase productivity and drive business value. An interactive AI use case registry gives your company full visibility into how people are using artificial intelligence right now. Pair that with capabilities building content in the form of tutorials, learning paths, and a use case library, and Superintelligent helps people inside your company show how they're getting value out of AI while providing resources for people to put that inspiration into action.

6:42The next three teams that sign up with 100 or more seats are going to get free embedded consulting. That's a process by which our super intelligent team sits with your organization, figures out the specific use cases that matter most to you, and helps actually ensure support for adoption of those use cases to drive real value. Go to bsuper.ai to learn more about this AI enablement network. And now back to the show. So this was everything that was swirling around this election. And yet nothing close to these doomsday scenarios came to pass. In fact, the things that did happen are almost notable for how low impact they were.

7:18There was a widely publicized Biden deepfake robocall that was placed to thousands of voters in New Hampshire in January. On the eve of the presidential primary, an AI-generated version of Biden urged Democrats to stay home instead of voting, basically saying, save your vote for the real election. That incident attracted national attention, was quickly investigated, and the culprit identified. The FCC imposed a million dollar fine on the robocall provider and proposed a six million dollar fine for the individual responsible. Hilariously, the person implicated said the entire incident hadn't been an attempt to influence the primary, but rather a stunt to raise awareness about political deepfakes.

7:52If so, kudos, because the amount of coverage relative to the amount of impact has been enormous. Beyond that, there's just been very little. We haven't seen armies of AI-powered bots overrun social media, a phenomenon which arguably played a larger role in 2016. We haven't seen scores of deepfake videos embarrassing candidates. Sure, there were a few fake celebrity endorsements created with AI. A generated image of Elton John wearing a pink MAGA jacket got attention for a little while. And maybe the biggest was when Donald Trump himself reposted a bunch of AI-generated images of women wearing Swifties for Trump shirts on his Truth social network.

8:27But that just ultimately prompted Taylor Swift to formally endorse the Harris campaign in order to fight the misinformation. And while some of those images were vaguely believable, the more common use of AI during the campaign has been in the form of political satire. There have been, of course, images and videos of Harris and Trump in romantic scenarios, a clearly AI-generated video of Elon Musk and Donald Trump dancing in the streets, Trump as a Pittsburgh Steeler, and of course it's pretty hard to argue that these are actually meant to mislead anyone. Now, we did more recently get a more serious version of the same category, when on Sunday, a MAGA-affiliated ex-account posted a video of Martin Luther King Jr.

9:05endorsing Trump, which quickly gathered 10 million views. The video was immediately denounced by King's daughter, who called it vile, fake, irresponsible, and not at all reflective of what my father would say. But even with that, there's still a huge gulf between something being vile, to use her words, and being misleading. In other words, there are very few people who believe Reverend King is still alive, and even fewer who think that he would endorse Trump. Speaking to the scandal, Rachel Toback, CEO of Social Proof Security, said, We're in kind of the throw spaghetti at the wall moment of politics and AI, where this intersection allows people to try new things for propaganda.

9:37I think we're going to see people try anything they can to influence the election. And yet even Toback announced that this video wasn't about being believable, but rather something that could prey on people's emotions. She said, We live in this strange world now where deepfakes are kind of used to say, but what if it were real, and using that in a partisan way. Still, the point of all of this is that it's very hard to argue that there has been a meaningful impact on these elections from generative AI. In their latest update from September, the Office of the Director of National Intelligence said, Generative AI is helping to improve and accelerate aspects of foreign influence operations, but thus far the office has not seen it revolutionize such operations.

10:12The risk to US elections from foreign AI-generated content depends on the ability of foreign actors to overcome restrictions built into many AI tools and remain undetected, develop their own sophisticated models, and strategically target and disseminate such content. They highlighted that foreign actors are largely focused on sowing division on issues like immigration, drug use, and abortion. The DOJ confirmed that they had disrupted a Russian operation around immigration in September. And so the question becomes, what is the reason that this threat has not materialized this year? Is it, one, that agencies have just been working all around the clock to counter the effort?

10:45Is it, two, that the guardrail set up by companies like Dali and MidJourney, which made it harder to create deepfakes, were actually working as deterrents? Or is it three just still too early? Sonny Gandhi, the VP of Political Affairs at Encode Justice said, I'm sure in another year or two the AI models will get better, so I'm pretty worried about what it will look like in 2026 and definitely 2028. On the one hand, it is undeniable that with better models and broader access to more powerful models, the threat of misinformation will continue. For example, although Sora got a lot of people nervous, in truth it hasn't been widely available yet.

11:23And the models that are are nowhere near as good as the total fidelity that would allow for a really detailed misinformation campaign. At the same time, it's hard not to read a statement like that and feel like it's just a massive moving of the goalposts for someone who wants to be scared. My base case had always been, coming into this election cycle, that I thought that the presence of generative AI was going to make people simply more skeptical of everything right from the get-go. Basically that in the era of generative AI, our assumption is going to be that we can't trust anything. If that's how it plays out, that will radically limit AI misinformation's ability to convince us of things, which is a good thing.

11:59But at the same time, it also creates its own consequences when no one believes anything. Obviously, one of the defining traits of this election could be voters not accepting the results of the election because our trust in institutions and officials is so low right now. Ultimately, where I land is that the issues that we face are much more about leadership, institutions, and trust in one another right now than they are about technology. It would be naive to think that AI couldn't exacerbate those things, make them worse. And it would be naive to think that as enhanced capabilities come online, there won't be more threats.

12:31But I do believe that this election is a good reminder that the histrionics we sometimes find ourselves in as it relates to the threat of generative AI are simply a lot messier and more complicated in practice. Now, of course, there is still a part of me, even as I record this at 9.40 a.m. on Election Day, that wonders if I won't have to come back on here tomorrow and discuss last-minute AI voter tampering or some other Election Day interference itself. But at the moment, it seems pretty clear that while the next administration might be defined by artificial intelligence, this election certainly wasn't.

13:06That's going to do it for today's AI Daily Brief. Appreciate you listening or watching as always. And until next time, peace.