“What a bunch of malarkey. You know the value of voting Democratic. Our votes count. It’s important that you save your vote for the November election. We’ll need your help in electing Democrats up and down the ticket. Voting this Tuesday only enables the Republicans in their quest to elect Donald Trump again. Your vote makes a difference in November, not this Tuesday”
That’s the transcript of a robocall of a voice sounding like US President Joseph Biden. And it was said to have been circulated just before New Hampshire’s Democratic primary election in January 2024. “Voice sounding like US President Joseph Biden” is key, because while it sounded like him, it wasn’t him: it was a deepfake and the idea may have been to make sure voters skip the New Hampshire primary. It’s a deepfake that may have been made using AI to create a convincing replication of a voice, which could help manipulate voters and people who may not be cognizant of the deepfake phenomenon.
So, with the advent of generative AI in late 2022, what could this mean for elections? Especially, since 2024 is considered *the* election year, not just any other election year. It’s said that more voters, than ever before in history, would vote in about 65 countries or even more this year. That’s about half the world’s population. So, the potential and risks of AI are something that should, probably, be looked at more closely in such an important election period.
So, could AI engage and mobilize voters? Or would it manipulate opinions, create more division and erode trust in institutions?
Especially when there are said to be deepfake videos of Bollywood performers making the rounds promoting certain political parties during campaigns for General Elections 2024 in India.
According to Sunil Nair, Executive Director at Concept Communications Ltd, “When I met Mr Narendra Modi during his tenure as Chief Minister of Gujarat for a project, that’s when I saw the advent of social media and realized the next war wouldn’t be fought on the battlefield, but on the Internet. And that was just a preview. Over time, there’s a profound change in how people think and their basic habits, in terms of how much time they spend on their handheld devices. A survey I came across outlines that people spend 3-4 hours daily consuming content on these devices. And it’s easier to engage with a simple Instagram Reel than to absorb the details of a mutual fund ad; people might remember the Reel but forget the nuances of the mutual fund ad. Many still don’t understand the mechanics of social media. Algorithms scrutinize content for political correctness and shape narratives, with AI playing a pivotal role in these new developments”.
Nair remarks, “On social media, it’s often impossible to distinguish between genuine accounts and imposters of public figures. And while deepfakes might seem like a recent phenomenon, they’ve been around for years, emerging prominently around 2017 or 2019. These are manipulations that have caused significant trouble and confusion. In India, deepfakes are prevalent in the political arena and the problem has become so normalized that people have stopped actively defending it. Deepfakes can drastically alter perceptions, mislead audiences and distort the truth about people, be it positive or negative”.
But, what about blue ticks? Social media platforms, like Meta and YouTube, are said to be labelling AI-generated content. Could that help?
“Blue ticks are now available at a cost, so blue ticks have no relevance today. So, that’s not going to be a differentiator”, opines Nair.
But, what about beyond deepfakes? What’s the impact of AI, then, on the election landscape?
A country, like India, is linguistically diverse. So, maybe, AI could be used to translate a political figure’s speeches in real-time without changing the voice. How about AI-driven real-time translation technology that could preserve their voice and style? Political figures may have a sense of disconnect with voters in countries, like India, which has multiple languages. If a political figure’s primary language is Hindi, it could be hard to connect with a Tamil Nadu base that only speaks Tamil. So, in this case, AI could revolutionize political communication. Yet, some contextual nuances and dialect differences could be missed and a mistranslation could lead to misunderstandings.
Nair declares, “The government operates on four key principles: Hindutva ideology, narrative-building beneficiaries, neo-nationalism and effective outreach to the lowest strata of society. Mastering outreach ensures a political figure’s message reaches every voter in their constituency. AI is integral to this strategy, with creators leveraging it to reach the last mile, given widespread smartphone usage and penetration in India. AI is highly effective in producing compelling content, videos and caricatures. We even run a virtual influencer program featuring ‘Naina,’ an AI superstar and we’re preparing to collaborate with brands. So, every point, in terms of messaging, can be done through AI”.
How about a digital twin of political figures to replicate their persona and voice? Maybe, this could be done by collating all their speeches, viewpoints and ideas. This could engage with constituents in real-time and create a sense of connection and camaraderie and help them understand the political figure’s stance on different issues. That could be a way to augment voter engagement and help engage with candidates more directly, maybe fostering a sense of connection that traditional campaigning may lack. On the other hand, a digital twin could be programmed maliciously to make false promises or spread disinformation.
“A digital twin can be reached to people through social media platforms, like Instagram or WhatsApp. If one can catch their attention for 10 or 20 seconds, they end up opening it”, states Nair.
So, could anything be done from a citizen’s perspective? If people are talking about certain issues and it shows up on their phone or on their feed, could that be curbed or stopped?
“There are voices, groups and companies looking at this option, because it’s taking away productive time. One may be doing something critical and when they see these kinds of things, attention is lost”, states Nair.
What about the actual voting mechanisms itself? Could AI play a role there?
Nair remarks, “It’s a little early to say whether AI could play a role there, so it’s a matter of time. It will play a role, not in the near future, but I think a couple of years down the line definitely. You may not even have to go out of your house to vote, especially when voting percentages have gone down”.
And down the line, could AI increase voter turnout?
Nair asserts, “There’s a case there. India is a large country, where the middle class thinks very differently from the lower middle class. If one is able to bring tech to one’s doorstep and allow them to vote without leaving their house, it could mean a new way of choosing one’s electorate”.
“If technology takes over, what will people do and where will they go from that point? The situation has both pros and cons. Over time, there may be a significant hue and cry. Unemployment stresses the economy and there’s no social security framework, so crime may rise, inflation may soar and the rich-poor divide may widen. And the wealthy will use AI to drive down costs and increase their wealth. But, what about the everyday worker? Yet, the world goes on. India aims to become the world’s fifth-largest economy and AI is going to play a crucial role”, quips Nair.