This article is adapted from Chapter 17 of the 2024 version of “How to Use the Internet to Change the World – and Win Elections”, now in its 11th edition. Download your copy of the definitive digital politics handbook today. Plenty of lessons for newcomers and experienced digital campaigners alike!
1. Artificial Intelligence
With the release of chatbots like ChatGPT, artificial intelligence first became a realistic tool for campaigns and advocates late in 2022. More chatbots and other AI platforms have since followed, accompanied by a great deal of hype. ChatGPT and its friends can seem almost magical when people first interact with them, but the current generation has limitations that should protect most us from a robot job takeover…for now.
A chatbot ultimately reflects our own language back to us, based on the correlations its machine-learning algorithms find among the billions of words of human writing it’s been fed. It’s not going to come up with an idea because it understands something in a new way, but because it’s put words together in a sequence that humans interpret as meaningful. But that’s okay! If a chatbot comes up with language or imagery that helps campaign staff or consultants do their jobs better, it’s a useful tool.
For instance, chatbots might help field organizers to “cut turf,” breaking down a voter file into lists of doors for volunteers to knock or phone numbers to call. Similarly, a data analyst could ask a bot to create a stack of spreadsheets from a big mass of advertising data, or look for patterns in polling crosstabs. Other communicators will use them to create ad variants or ideas for content, often after “training” the AI by exposing it to samples. Behind the scenes, AI tools may help our us reach the right voters cost-effectively, since they’ll be baked into platforms from social media to advertising to email deliverability.
Many users don’t realize, though, that the information they feed into public chatbots can become part of a bot’s larger data set. It may appear in the results the platform shows to someone else! Political actors should be quite wary of what they enter into a chatbot’s prompt, taking pains to anonymize or otherwise obscure the data. For instance, if I’m asking a chatbot to draft a statement of work for a new project, the kind of task at which these tools typically excel, I’ll fake enough of the details to keep the client’s secrets.
In part to avoid those problems, companies in the political space are creating AI tools based on proprietary data sets and licensed AI technology rather than public chatbots. We mentioned Quiller in the fundraising chapter and SoSha in the social-media section, two companies that help left-leaning and Democratic campaigns and organizations create content based on libraries of curated content. I’ve also spoken recently with companies that use licensed AI engines for other campaign tasks like opposition research and preparing press clips.
In a broader sense, campaigns and organizations may already be using AI without realizing it. Like the AI apps that help people edit photos on the fly on their phones, common image and video tools like Adobe Photoshop and Premiere now include AI assistance. If you’re using a video created in the last year or so, the producers may have had help from AI when they put it together. Cases like this one complicate attempts to label “AI-generated” content on social media or to limit it in political contexts. When everything gets an AI-assist, what becomes the threshold for disclosure?
Of course, some companies will take a standard tech product, sprinkle a little AI on it and turn on the marketing machine. Already, my inbox regularly features pitches from AI-driven content marketing companies, sales-lead generation platforms and the like, some of which may not be as legitimate as others. Buyer beware!
Another pitfall? Chatbots look at patterns in the ways humans use words, not at how those words relate to the world itself. AI can lie! Today’s chatbots routinely make up “facts” when they’re asked to create content, and one suspects they have little instinct for the nuances of politics. AI can turn out content, but any campaigner who feeds a chatbot’s words or pictures unedited into the world is asking for trouble. And those words can be pretty terrible, since a basic request to “write me a fundraising email” without context or detail can turn out some wretched results.
Similarly, an AI’s analysis of your marketing data will functionally come out of a “black box” — you’ll see the results, but you won’t have any real idea about why the system arrived at the conclusions it did. AI does not give us easy ways to check its work! So be sure to build in points in the process that allow you compare an AI’s output with information you know is correct. You don’t want to make decisions based on bad data any more than a comms person wants to put out a press release apparently from an alternate universe.
Actually, bad actors WANT to spread tales from alternate universes! AI can churn out disinformation as easily as it can create any other writing or imagery, and as we have seen repeatedly in recent years, lies can spread extremely quickly online. AI neither knows nor cares how you’ll use its work product, and it could turn out to be the most powerful digital trickery tool of all times. Some of us have been talking about the potential danger of AI-created fake video for a while, but now we’re already dealing with the fallout in real time. It’s only going to get worse.
Despite these risks, campaigns and other political communicators will surely employ AI for a bewildering array of tasks this year. Armed with the right queries and the right caveats, and after a little experimentation, campaigns and advocacy organizations can and will use AI tools to get more done in less time in 2024. Watch this space — AI will almost certainly evolve more rapidly than any other slice of campaign technology in the next few months.
2. Social Media
Our relationship with social media? Complicated. The companies see themselves as a force for good: Facebook founder Mark Zuckerberg, for one, clearly feels that connecting people in new ways is inherently positive. But as we’ve seen many times in this book, digital tools are blind to their ultimate use. In practice, people have found it just as easy to use Zuckerberg’s baby to organize a riot — or a massacre — as a happy hour.
Platforms built to value engagement over truth naturally favor the lurid and the extreme, and conspiracy theories and disinformation thrive without some mechanism to filter content. How many people found QAnon and anti-vaccine videos because Facebook or YouTube recommended them? Millions, and not just in the United States. The platforms WANT to be the center of our digital universes, but they haven’t put in the time and money to tackle the responsibilities that accompany that status. Case in point: Elon Musk’s gutting of Twitter/X’s moderation practices right before our eyes.
Russian social-media manipulation set off SOME alarms after 2016, though I doubt their work shifted many votes in the U.S. elections. The real enemy was us all along: anti-vaxxers and Q got a lot of attention, but they’re just two of America’s many homegrown social-media horrors. In a real way, the 2021 Capitol riot was the natural outcome of a digital environment that locks us in our personal information bubbles.
Lies spread fast with an assist from prominent voices, and I was not remotely surprised to find that political misinformation plummeted on Twitter when it first banned Donald Trump. Though lies rarely require advertising to reach gullible ears, tech companies’ “solutions” have usually been limited to placing hurdles in the path of legitimate political advertisers or banning paid political content entirely. In the process, they’ve made it HARDER for campaigns and activists to respond to smears and falsehoods.
The answers? I don’t have easy ones. For a start, perhaps Facebook and company could put the kind of energy into slowing down conspiracy-theory superspreaders that they do into policing copyright violations. I’m as much of a free-speech purist as you’re likely to find, but we don’t generally put up with incitement to violence when we encounter it in real life. Why should we allow companies to profit from hate and fear in the virtual space? Restrictions on the spread of “stop the steal” content are not ideal, but these lies also killed people in the physical world. They are far from alone.
With all that in mind, you might not be surprised that political social media somewhat up in the air in 2024. Besides incitement and active disinformation campaigns, digital activists have to deal with the lingering fallout from ongoing data-breaches, Russian election-hacking and Cambridge Analytics data-scraping, including disclosure and verification requirements that can put a serious crimp into a campaign’s plans if they don’t account for them in time.
More fundamentally, political communicators face the same problem as everyone else trying to reach people through social media: getting our content in front of their eyeballs in the first place. On Facebook, for example, most of our followers will never see our content unless we pay for the privilege. This year, the company has put even more obstacles in our path, by limiting the visibility of political content on Instagram as well as Facebook itself. TikTok doesn’t penalize organic political content that we know of, but it also does not allow campaigns or outside groups to advertise on behalf of a candidate. Only videos that generally encourage viewers to vote can run as ads. Meanwhile, YouTube ads can be relatively cheap, but political advertisers’ options are strictly limited. No voter-file targeting on Google properties!
When confronted with obstacles like these in the past, campaigns and advocacy organizations have tried to rely on a mix of compelling content, active supporters and paid promotion to get their messages in front of the right voters. Smart campaigns also spot opportunities to take advantage of spontaneous social enthusiasm. Elizabeth Warren’s campaign didn’t create those now-legendary selfie lines out of the blue — they capitalized on a practice her passionate supporters started themselves. Still, we can’t fully depend on zeal to translate into persuasion, since people can only be persuaded by what they can see. At least on Facebook and Instagram, potential supporters may not encounter much of the political content our followers share.
As we’ve seen in earlier chapters, campaigns and advocacy organizations can try to extend their social-media reach by employing influencers, content creators and others with large (or highly relevant) followings on Instagram, TikTok or some other strategic platform. Joe Biden’s 2020 campaign centered outreach to content creators in its social-media and young-voter programs, and they rewarded him by making TikTok his second-biggest video channel, even though the campaign had no formal TikTok presence of its own at the time. Biden’s 2024 campaign is taking a similar approach, already holding events specifically for influencers.
Influencers may be relatively easy for a presidential campaign to recruit, but what about someone running for state representative or trying to become the mayor of a small city? The big question about influencer outreach is whether it can scale down-ballot, particularly whether it can replace the utility of organic content and political ads for campaigns that aren’t as much in the spotlight.
3. Data-Driven Voter Targeting
Complicating the picture even more? Many voters are effectively hiding from political campaigns. Traditional American political advertising practices can’t always handle a world in which voters abandon broadcast media in favor of Netflix, Hulu, Spotify, YouTube and TikTok. Meanwhile, the death of land lines and the ubiquity of cell phones complicate phone-banking, robocalling and polling. As far back as 2012, the Obama campaign could only find phone numbers for about half of the voters aged 18-30 on its target list, and many had moved so often that their physical addresses were dead as well. No matter how good your data model is, it won’t help if you can’t actually reach the people you need. Fortunately for those of us trying to influence and mobilize the public, digital options can fill the gap.
Campaigns can now employ a wide array of digital targeting techniques to contact defined groups of voters via “addressable” communications channels. As we saw earlier, programmatic digital and streaming-media ads can be targeted via a voter file to reach specific households with messages designed just for them. Powerful, assuming that your voter model is correct in the first place. This level of precision has also entered the traditional TV world as well.
Similarly, a Facebook Custom Audience can help you reach individual people with messages optimized to move them. As Trump found in 2016, Facebook’s ability to “retarget” people who interact with specific content or who visit your website also provide ways to convert someone’s passing interest into a durable connection. Your own email list constitutes a target in waiting, since you can break it into specific segments based on supporters’ giving history, past participation in online actions, location or social media activity.
Keeping track of all that data requires systems that can talk to each other. Data integration isn’t all that sexy — unless you really, really get into mingling ones and zeroes — but it’s a secret hero of digital politics. The tools have powerful potential, but all the data in the world won’t help you if you can’t organize it, connect it and put it to work.
Think of all the data sources campaigns can access: fundraising lists, voter files, grassroots contacts, commercial information, demographic data and much more. Making all those pieces talk to each other is a major task, one that even seriously data-focused campaigns can’t solve perfectly. The most effective data-driven campaigns will be ones that 1) understand the data they have or can have, 2) connect different data “silos” to create integrated systems, and 3) use that data to make better decisions about how, what and when to communicate with voters. Ideally, the system will also incorporate feedback loops, so that information gathered via one source can affect the rest. If someone signs up on the email list from a given neighborhood, the local field organizer should know about it!
Money, People or Both?
The fact that you CAN buy targeted ads in bulk means that they become a button you can push: insert the check and the ads pop out. Will they also encourage a manipulative mindset, a return to the attitudes of the era when broadcast TV advertising dominated politics? The ease of targeted digital ads could encourage political professionals to treat the voters like clay to be shaped — as bystanders, not actors themselves.
My guess: we’ll see a mix of the two strategies, often in the same campaign. Done right, online ads and field organizing work together, reaching voters with consistent messaging across many points of contact. Voters not reachable digitally may be easier to find in person — or vice versa. But I also suspect that we’ll continue to see a tension between ad-targeters and those with a grassroots mindsets…particularly when it comes time to allocate campaign budgets. Note that Bernie Sanders may have been famous for his people-powered campaigns, but his team also spent heavily on digital ads in both of his attempts to win the White House.
4. Bots, Trolls, Hacks, Fakes and other Outside Interventions
If we had any illusions about Bad Hombres in U.S. politics, the last few years should have broken them. In 2016, Russian hackers and bot-nets did their best to heighten the discord in our politics, dancing on the raw nerves in our culture to pit group against group, sometimes literally. After this experience, it’s obvious that any candidate can be on the receiving end of lies spread online, whether it’s amplified by a bot-net or your crazy uncle. In this environment, every campaign should plan for rapid response against a digital smear. Pro Tips: mobilize your supporters to speak on your behalf, and be sure not to repeat the lies as you fight them. And if you thought fake news and Facebook data breaches were bad, we no longer have to wait for AI-created fake video and audio — it’s here.
In this environment, campaign cybersecurity cannot be an afterthought. Every campaign, activist organization and individual advocate needs to think about basic measures like secure passwords, two-factor authentication and virtual private networking apps. Also take a look at third-party tools designed specifically for campaigns, journalists and interest groups.
Of course, plenty of “legitimate” outside groups also try to influence elections, including SuperPACs and 501(c)4s. Most independent expenditure groups have poured the bulk of their money into TV in the past, but look for outside groups to invest more heavily in digital advertising these days, in part to pick off the shakiest parts of their opponents’ coalitions. For inspiration, note the integrated online/offline campaign waged by a SuperPAC that helped beat David Vitter and put a Democrat in the Louisiana governor’s mansion late in 2015. Or, look at targeted digital ads flying during the the Great Obamacare Battle of 2017, Congressional special elections in the Trump era and the Georgia Senate double runoff early in 2021. Finally, liberal activist communities like Daily Kos and many thousands of individual small-dollar donors have fueled insurgent Democratic campaigns and organizations, at times raising hundreds of thousands of dollars almost overnight. Unfortunately for Dems, however, 2020 exposed the limits of using political money to buy electoral love.
If you’re a local candidate, you might not have big outside groups intervening in your election. You might have upset local residents intervening instead! Either way, your own supporters are likely to be your best allies. Encourage them to be your ambassadors in their own social circles. Many of them will be eager to do so.
5. The People Will Not be Ignored
As we’ve seen throughout this book, candidates and advocacy organizations inhabit a political environment profoundly changed by the rise of a newly engaged electorate. As the massive voter turnout in recent U.S. elections has shown, millions of people are fed up with the way things are and are fired up to change it. No longer content to watch TV ads and vote as ordered, the people want in on the game. As I wrote when Trump was still in the White House,
Think of the Parkland students crossing the country to work for gun control, or the thousands of Indivisible chapters that have sprung up spontaneously in all fifty states, or the Democratic women organizing in secret in deep Red America. The Right gets in on the action sometimes, as in that border-wall crowdfunding campaign, but the real passion has been on the Democrats’ side since Trump went to the White House. If I were running for president, one of my priorities would be to create a system to harness it.
Not just in the United States, either, since protesters have challenged political authority around the world in recent years. The trick for campaigns and activists is to understand this enthusiasm and channel it toward positive ends. Crucially, this kind of people-powered politics treats supporters as an active part of the process and a major outreach channel in themselves. Voters are people with whom we have a conversation — they’re more than passive receptacles for our advertising.
In 2016, Bernie Sanders’s volunteers famously helped his campaign with everything from technology to field organizing, donating millions along the way and bringing him closer to the Democratic nomination than almost anyone expected. Meanwhile, Trump supporters did their best to retweet the future president to the White House (with a little help from Moscow), and their grassroots donations basically paid for the last two months of his victorious campaign. After Trump’s inauguration, Left-leaning activists in the “Resistance” self-organized to a remarkable degree, bringing a new energy to progressive causes. Their chief rivals? The members of Trump’s “base”, whose passions he still stokes through furious and fuming missives posted online. From racial justice to immigration policy to abortion, the conflict between these competing visions will power elections in the U.S. in 2024 and beyond.
With EVERYTHING seemingly on the line in each election for school board on up, I don’t see the American public lapsing into apathy any time soon. Burnout may become a more serious factor if we don’t relax at least a little bit in the years to come, but digital tools simply make it too easy for would-be activists to act on their dreams and demands. Their wide reach means that a single person with the right ideas presented well can connect with people far beyond any one person’s limited social circles. The cat is out of the proverbial bag — and he’s gone online.
Everything You Just Read is Wrong
Well, maybe it’s not ALL wrong, but plenty will happen in 2024 and beyond that we can’t predict. Some trends that look significant now may not play out that way in the end, and strategies that work well one month can yield eyeball-melting failure the next. That’s what keeps this business fun! We’ll revisit these trends on Epolitics.com regularly in the months to come.