The era of AI in politics is here. With the right safeguards, it can benefit society

This column was written entirely by me, with no involvement or assistance from artificial intelligence whatsoever.

The same, however, cannot be said of robocalls made January in the New Hampshire primary that appeared to be from President Joe Biden; nor of recent audio that seemed to capture the voice of Manhattan Democratic Party leader Keith Wright. Both of these high-profile incidents were deepfakes — phony recordings generated by artificial intelligence designed to sound authentic but in reality are fabricated. And whether it’s an imposter President Biden or a counterfeit Keith Wright, both episodes portend the threat posed to our elections and our democracy by bad faith actors skilled at manipulating AI.   

The era of artificial intelligence in politics is officially here, whether we like it or not, and the truth is the current regulatory landscape is insufficient to prevent the dangers of widespread fraud and disinformation. This will be the dominant tech story of the 2024 election cycle, up and down the ballot, and the consequences of how we respond now will reverberate for years.

But the choice before us is a simple one: Do nothing and let deception further corrupt our politics, dupe voters and imperil free societies; or demand real regulations that not only defend the voting public against fraud — but allow us to unleash the positive benefits that AI can offer us all.

Artificial intelligence possesses enormous potential to democratize politics by transforming voter contact with nuanced communication at massive scale, and reducing dependence on big donors. It’s exciting to stand on the precipice of a technological breakthrough.

But, without proper safeguards, we will lose organic control over what is an inorganic political technology. This makes it much more likely that those with dark designs can undermine democracy by wielding AI as a disinformation weapon. 

By next year, most Americans will have already spoken to or interacted with generative AI at some point; most of these same Americans won’t even know it. It becomes easy to envision scenarios where AI is exploited to spread misleading information about candidates or election results at massive speed and scale, whip a mob into a frenzy, and then direct that horde to launch another Jan. 6-style attack. Or tell voters they don’t need to vote on the actual election day. Or impersonate a candidate saying outrageous, vile things that repel voters. Or be leveraged by a hostile foreign government to interfere with our electoral process to deliver results that serve their interests at the expense of our own.

Preventing these nightmares must be the mission for all of us who want to see AI reach its spectacular positive potential, and serve as a healthy and productive force in our public discourse. What’s needed are clear laws that establish the rules of the road for responsible use, while  maintaining organic control so we know what material comes from people, and which comes from machines. 

Transparency remains the decisive factor to ensure this control. Importantly, the risks should not open the door to reactionary extremism in the other direction. Calls for outright bans are misguided and shortsighted — and impossible to enforce. Used properly, AI is an empowering tool for political campaigns and movements. The promise of candidates and public officials engaging in real, interactive one-on-one discussions with the people they represent in a personal way is revolutionary. 

Imagine a more personal way for campaigns to connect with voters — instead of blanketing TV airwaves and social media feeds with ads that just shout at voters. Picture underdog campaigns with little funding able to level the playing field against political opponents flush with cash from wealthy donors. Visualize underrepresented and marginalized people historically shut out of the political process able to compete equally with the entrenched power centers. And envision a government able to interact at scale and speed with constituents in real two-way communication about signing your kid up for school programs; learning about closures during a snowstorm; or how to get to your destination when subway lines are out of action.

This is the promise of AI that is so exciting: not just playing tape at voters, but connecting with them, learning what’s important to them and providing real time answers and information to their questions that no voter interaction strategy or technology available today can provide.

AI has the potential for mammoth positive societal change. The key is approaching this new reality with eyes wide open, because there is no putting the genie back in the bottle. That means marching deliberately and clear-eyed into the future, not sleepwalking or stumbling into it, nor vainly trying to prevent it either.

If we get this right, with meaningful regulation anchored to transparency, AI will truly serve the betterment of humanity. And the next time you read a political op-ed about campaigns or government, you can be confident that the words come from the people whose names appear alongside it (like this one), and not from nefarious AI spinning toxic political lies with zero accountability.

Ilya Mouzykantskii is the CEO of Civox, the world’s first voice AI political company.