what election officials call “ trusted voices .” These voices are credible people inside communities , online or otherwise , who will say “ not so fast ,” defend the democratic process , and hopefully be heard before passionate reactions trample facts or run amok .
“ All the tools we need to instill confidence in our elections exist ,” Benson said at the summit . “ We just have to get them — not just in the hands of trusted voices , but then communicate effectively to the people who need to hear them .”
But Benson , whose hopes echo those heard from election officials across the country , was putting on a brave face . She soon told the audience that reining in runaway rumors was not easy — and said officials needed help .
“ We need assistance in developing trusted voices ,” Benson said . “ Second , we need assistance , particularly from tech companies in both identifying false information and removing it . We know we ’ re farther away from that than we have ever been in the evolution of social media over the last several years . But at the same time , where artificial intelligence is going to be used to exponentially increase both the impact and reach of misinformation , we need partners in the tech industry to help us minimize the impact and rapidly mitigate any harm .”
Still , she pointed to some progress . Michigan ’ s legislature , like a handful of states , recently passed laws requiring disclosure of any AI-created content in election communications . It has criminalized using deep fake imagery or forged voices . Congress , in contrast , has yet to act — and probably won ’ t do so before 2024 . The White House recently issued executive orders that impose restrictions on AI ’ s use . While the large platforms will likely comply with labeling and other requirements , 2024 ’ s torrid political campaigns are likely to be less restrained .
And so , Benson and others have tried to be as concrete as possible in naming plausible threats . In November , she told the U . S . Senate ’ s AI Caucus they should “ expect ” AI to be used “ to divide , deceive , and demobilize voters throughout our country over the next year .” Those scenarios included creating false non-English messages for minority voters , automating the harassment of election officials , and generating error-ridden analyses to purge voters .
But countering propaganda remains a steep climb . Even before AI ’ s arrival , the nuts and bolts of elections were an ideal target because the public has little idea of how elections are run — there is little prior knowledge to reject made-up claims and discern the truth . Moreover , political passions peak in presidential years , and every election cycle has a handful of errors by workers who set up the devices that check in voters , cast votes , and count ballots . This November ’ s election was no exception , according to this tally of administrative errors by ElectionLine . org . After Trump ’ s 2020 loss , operational mistakes like these — which caused some votes to be incorrectly counted before the error was found and corrected — were misinterpreted , mischaracterized , and became fodder for his stolen election tropes . Much the same scripts surfaced after this fall ’ s snafus .
As crisis informatics researchers like Starbird have confirmed , doubt is more easily provoked than trust . Nor do labels and fact-checking alone change minds . At the cyber summit , Raffi Krikorian , the CTO at the Emerson Collective , a charitable foundation — who held that role for the Democratic National Committee during the last presidential election — made the same point .
“ Our election system , in a lot of ways , is built on trust — like we trust a lot of the portions of the mechanics will work together ,” he said . “ There ’ s a lot of humans involved . It ’ s not necessarily all codified . And so , [ if there is ] any break in the system , it ’ s really easy to lose trust in the entire system .”
Krikorian , who hosts the Collective ’ s “ Technically Optimistic ” podcast , was worried about 2024 . Turning to Benson , he said that his research about “ where people [ in Michigan ] are actually getting their information ,” shows that it is increasingly from smaller and more obscure online platforms . Others at the summit noted that people were turning to encrypted channels like WhatsApp for political information . In other words , the propaganda pathways appear to be shifting .
“ We ’ re actually , in some ways , spending our time maybe in the wrong places ,” Krikorian said . “ We need to be spending our time on these other smaller and up-and-coming platforms , that also don ’ t have the staff , don ’ t have the energy , don ’ t have the people and the resources needed in order to make sure that they ’ re secure in the process .”
Nonetheless , the remedy that Benson kept returning to was locally trusted voices . For example , her office has stationed observers within 5 miles of every voting site to investigate any problem or claim . She said that she has been engaging “ faith leaders , sports leaders , business leaders , and community leaders about the truth about how to participate and how to trust our elections .”
“ AI is a new technology , but the solution is an old one — it ’ s about developing trusted voices that people can turn to get accurate information ,” Benson said . “ All of this , everything we ’ re talking about , is about deceiving people , deceiving voters who then act on that deception .”
But rumors , misperceptions , mischaracterizations , deceptions , lies , and violence in politics are as old as America itself , as political historians like Heather Cox Richardson have noted . There is no simple or single solution , said Starbird . Nonetheless , she ended her Stanford address with a “ call to action ” urging everyone to redouble their efforts in 2024 .
“ We ’ re not going to solve the problem with misinformation , disinformation , manipulation … with one new label , or a new educational initiative , or a new research program ,” she told trust and safety researchers . “ It ’ s going to have to be all of the above … It ’ s going to be all these different things coming at it from different directions .”
The Trial Lawyer x 69