How to Prevent AI from Tricking Us in the 2024 Election
Deep-Fakery Coming to a Polling Location Near You?
By Mark Joseph Mongilutz
We know it will be a problem eventually, and if recent Axios reporting is to be taken seriously, it might already be just that. This world of shifting sands and increasingly powerful AI technologies, all evolving at a pace beyond the population’s ability to track, will one day encroach on the sacred territory of democratic processes, of not merely your right to cast a ballot, but on your ability to do so at a location and in a way that ensures it will be counted. In other words, AI shenanigans could soon be viably deployed as confusion-generating tactics amongst an American populace that not only wants to believe the election guidance it receives is accurate, but has long been able to safely maintain that belief. Emphasis on “has”.
Concerned parties (ideally, every voting-age American) should rest assured that there are precautionary measures all of us can take in counteracting deep-fake efforts by tech-savvy, if conscienceless, bad actors. Before getting to those, I’ll advance here a plea that this very real problem not be drawn-and-quartered strictly along the lines of Right/Left political discourse. Like so many matters having to do with AI and the hyper-accelerated technologizing of our world, we would do well at the policy level in treating this with commensurate social fervor, but also with a mindset of political agnosticism – if it ends up falling down along Red vs Blue lines, the path towards engineering effective safeguards will be bogged down in “(R) said, (D) said” foolishness for years, leaving none of us any the better once the digital dust settles. If we view as essential our own right to vote, we must by extension view as essential the rights of our every countryman to do the same – that means insulating the voting process itself from AI-achieved distortions for everyone.
That all said, what are we really looking out for and what should you do to ensure your democratic rights aren’t impinged on by convincing AI tools?
There is an asymmetrical aspect to AI-powered technologies, one that precludes simplistic characterization. We can’t predict with specificity how and from where every tendril of misinformation will slither into the public arena, only that they will do so in various forms and that (at the risk of over-dramatizing the phenomenon) your instincts will often be the best first-line defense.
With AI at the fingertips of millions, deep-fakes are going to proliferate in politics and elsewhere. But as they do, so too I trust will our collective “uncanny valley” detector. This is the term in robotics for that feeling of “it’s not quite real” when you find yourself looking at an effort to simulate the organic. Over-abundant CGI in film has given us plenty of opportunity to refine our “that’s a fake” reflexes. At some point, if not many points, you will be presented with video footage or audio recording of your favorite political candidate saying something you can’t quite reconcile with who you know them to be. This could also happen in reverse, wherein people are quite happy with what the footage or imagery seem to suggest is happening. Many were apparently fooled by “photographs” of Trump fleeing his would-be arresters after the Bragg indictment came down.
Craftily rendered as some of them were, they were far from believable and rather easy to disprove. But there were also tens of millions of Trump supporters who had every interest in seeing to the disproving. What about local candidates whose support is limited to town- or district-level populations? In those cases, trust your reflexive interpretation and dig more deeply into whatever’s been reported (visit multiple news sources, check campaign home pages, seek out corroboration; more on that below).
There are also various tells you can look for, many good ones provided by the folks at Medium, who really do know what they’re talking about on this topic. The line between what’s plainly real and what is plainly fake is fading, but it will never permanently vanish. BS detectors are not all equal, but we do all have them – if you know something isn’t right, there are likely good reasons for it.
Don’t entrench yourself in the information line of a single news source, even one you’ve grown to cherish and rely upon. Drawing from a variegated crop of outlets will allow you, through cross-referencing, to more assuredly identify suspicious campaign claims, inaccurate voting/election guidance, and outright distortions of a candidate’s words, image, or positions.
If the deep-fakers will count on anything, it will be narrow media diets on part of those they’re aiming to deceive. If they can infiltrate relatively limited information corridors, they can pollute an audience’s understanding of actual events and bet that corrections to the record will be slow in materializing. Keep your media diet (like your culinary diet) diverse and you’ll be more likely to swiftly catch on to intentional efforts at misinforming you, particularly when it comes to manipulations of highly visible politicians and candidates, whose actual public profiles and real statements will be sufficiently monitored and represented to give any fakery a short shelf life in the media sphere.
Resist the partisan impulse. You’ll be tempted to believe the worst about those you plan on voting against. But don’t let that temptation lead you astray in discerning what’s objectively true from what’s harmfully false. The ability to mislead convincingly and broadly is a real problem for the republic not just because one or the other party stands to benefit or suffer, but because widespread deception harms us all on a human level. If you see a candidate whose political positions you revile being subjected to AI-enabled deep-fakery, be as quick to identify it as you would if your own candidate were being thusly attacked.
Make use of publicly available election resources, both state-level and federal. With intentionally contradictory and deceptive AI-generated nonsense sure to proliferate across the political landscape in coming years, visiting official election and commission pages will be useful in helping you discern truth from fiction for the likes of election dates, voting locations, and ballot-casting provisions. Be especially mindful of sources when receiving unexpected, surprising, or apparently outlandish news. The unexpected, the surprising, and the outlandish can occur in any election cycle, but learning of such things via, say, a stray TikTok clip is reason enough to seek additional information, the absence of which will usually (though not always) be reason enough to disregard the “news” in question.
There’s good reason for skepticism these days, but don’t let hysteria dominate. As has always been true of elections, there will be many who feel it appropriate, even necessary, to ensure victory for their side at any cost. Unfortunately, the costs for disseminating false-but-convincing material are dropping by the day. On the positive side of the ledger, combating falsehoods is as easy as learning how (and choosing) to detect them. But another option available to you and everyone is to get more involved in politics. Direct engagement with a campaign at the “grass roots” level will render the political world more substantive, something you can know viscerally, not strictly as a collection of flashing images and alarming headlines (see above).
A major source of technology’s hold on society these days is that so much of our knowledge of the world comes not through interconnectedness with it, but through visual simulations of it. When we step away from screen culture (admittedly hard to do, particularly when one writes for a living) we can wipe our eyes and properly realize that we’ve been responding to illusions rather than taking in life and the world itself. One way to do that is to volunteer for a local campaign, or at your nearby voting location. If politics is of real importance to your life, it should be a real presence in your life. After all, the deep-fakes only go as far as the glow of a phone screen, a computer monitor, or a television. We’re not yet at the physical deep-faking point – God help us when we are.
Stay skeptical, seek out information from the source, and prime your fakery-detection senses. We’re all operating on instinct to one degree or another; when confronted with thorough falsehoods (especially those that play to our political inclinations), we’ll sometimes need to trust those instincts more than the images on display. It won’t be your eyes lying to you, just something lying to your eyes… and all too well.
Note: the opinions expressed herein are those of Mark Joseph Mongilutz only and not hosts Chuck Warren and Sam Stone or Breaking Battlegrounds’ staff.
