2024 Polls: AI Fake News Will Damage Democracy Forever

2024 Polls: AI Fake News Will Damage Democracy Forever

Disinformation spreaders need not have any moral compass; fact-checkers, on the other hand, are accountable to the truth

Sachin KalbagUpdated: Thursday, October 26, 2023, 12:12 AM IST
article-image
2024 Polls: AI Fake News Will Damage Democracy Forever | representative pic

I am calling it now. In 2024, when several countries around the world — including India and the US — are scheduled to go to the polls, disinformation and fake news generated by artificial intelligence software will bring democratic systems to their knees. The damage is going to be so severe that, in the worst-case scenario, we will have reached a point of no return. Even if we somehow manage to halt the fake news juggernaut (unlikely, but let’s say we do), it will take decades to go back to any kind of normalcy.

We live in an age where ChatGPT, the revolutionary generative AI tool, has become old news in less than a year after its much-heralded launch last November. Its newer version, GPT4, was launched in March this year, and it is already just one of the thousands of such pieces of software.

Generative AI tools have been used and misused to create hundreds and thousands of pieces of disinformation and fake news around the world just this year; some so severe that they are responsible for scores of deaths such as in the ongoing Israel-Hamas war. In India, the civil unrest in Manipur started in May this year primarily after an out-of-context video of a woman’s murder went viral.

The next phase of disinformation and fake news will involve mass creation of such content. We have a tool just for that. A few months ago, a new AI-based video editing software entered the mass market. It has the capability of not only putting a voice over your video in eight languages, but it can also lip-sync any text to almost 100 per cent accuracy. It’s called HeyGen, and it has exploded into global consciousness within weeks. A HeyGen not just sounds like you, it accurately mimics you so that it also looks like you.

One language in its database is Hindi. The other seven are English, French, Portuguese, Spanish, German, Italian and Polish.

This is how it works. If you record a video of, say, you wishing your parents on their wedding anniversary, you can use HeyGen to not only translate your audio into any of the eight languages but also lip-sync in that language. In terms of technological achievement, this is a triumph. Imagine you sending a message to your long-lost friend from a European university in her language? Or, two transcontinental company executives talking to each other and signing multimillion dollar deals without the need for a bridge language or an interpreter? Or, your company generating a video advert for a region you want to expand to, but have no local expertise in? The possibilities are endless.

But here’s the thing. No political party is going to use HeyGen or similar software to greet you in your language or send a ‘Good Morning’ message. Instead — and I am certain about this — social media teams of political parties are going to use this to record manipulative and patently false videos of their rivals.

HeyGen and its contemporaries are so easy to use that you do not need any kind of technical knowhow.

According to HeyGen CEO and co-founder Joshua Xu, users can record a short video of themselves speaking, which trains an AI avatar clone. You then type any (yes, any) script and it generates a video of the resulting avatar delivering those lines in their own voice. There is no filming required.

The AI bit of the software is so strong that it can put in music, background scenes, graphics, etc. to produce a high-resolution video from just that simple text script. You can literally make Rahul Gandhi or Narendra Modi say anything.

Here’s a thought experiment: Imagine you belong to the dirty tricks department of a political party, and you are charged with creating a controversy — say, a sex scandal or a massive corruption case — involving your rival. You can now put any text into the software, and generate any high-resolution video on the basis of just one archival clip.

What could go wrong, right?

The truth is, no one can halt the evolution of technology. In March this year, soon after the GPT4 launched, more than a thousand leading names in business, technology and science — including the world’s wealthiest man Elon Musk — published an open letter asking the world to pause further developing the AI tools. They said that further development should be halted for another six months because of “potential risks to society and humanity.” The letter was issued by the Pennsylvania, US-based non-profit Future of Life Institute.

The letter called for a pause on AI development until safety protocols were developed, implemented and audited by independent experts. The communiqué said, “Powerful AI systems should be developed only once we are confident that their effects will be positive and their risks will be manageable.”

The letter also explained how AI software can pose risks to society and civilisation in the form of economic and political disruptions. The influential signatories, including those who pioneered AI tech, called on developers to work with policymakers on governance and regulatory authorities.

Well, no one cared; and frankly, it is pointless to suggest technology will regulate itself.

So, why should we be scared?

AI-generated disinformation and fake news videos have begun weakening our social fabric. This is not even news any more. Until now, fact-checkers and aware netizens would spot fakes on the basis of lip-sync issues or just that the video ‘looked’ fake. But the next phase of this disinformation project involving HeyGen or HeyGen-like software will make it much, much tougher to spot fakes.

Fact-checkers will still be able to give us answers and source it back to an original video or a document that may say the exact opposite, but by then it would be too late. Fake news travels around the world even before the truth can tie its shoelaces.

Disinformation spreaders need not have any moral compass; fact-checkers, on the other hand, are accountable to the truth.

When the nation’s political discourse is based on disinformation and fake news — the election frenzy next year is certain to exacerbate this — it is but one step to the weakening and eventual destruction of democratic institutions. And since we cannot stop the march of technology (no one listens to even the world’s richest person, after all), we may well have reached a point of no return.

Sachin Kalbag, a journalist and a podcaster, is Senior Fellow at Takshashila Institution, a Bengaluru-based non-partisan think tank. He can be reached at sachin@takshashila.org.in, and tweets at @SachinKalbag

RECENT STORIES

Shahpur Kandi Project Will Benefit Farmers As India Takes Daring Decision To Stop Water From Barrage...

Shahpur Kandi Project Will Benefit Farmers As India Takes Daring Decision To Stop Water From Barrage...

Editorial: Who Is Responsible For The Blasts?

Editorial: Who Is Responsible For The Blasts?

Editorial: Masala Thriller Okayed, But What About Its Impact On Real-Life Case?

Editorial: Masala Thriller Okayed, But What About Its Impact On Real-Life Case?

Analysis: GDP Second Advance Estimates Encouraging

Analysis: GDP Second Advance Estimates Encouraging

Analysis: Tackling The Forces Fuelling Inflation

Analysis: Tackling The Forces Fuelling Inflation