Experts concerned: Is deepfake turning into a weapon of mass deception?

Approximately two billion voters will head to polling stations this year to cast their ballots for their preferred representatives and political leaders. With major elections taking place in numerous countries, including the US, UK, and India (as well as for the European Parliament), this has the potential to alter the political landscape and the direction of geopolitics for the next few years and beyond.

As fake images, audio, and video clips of real people become more widespread, the possibility of AI-powered disinformation is causing growing concern.

Fake news has dominated election-related headlines since it became a major story during the 2016 White House race.

Eight years later, there is arguably a bigger threat: a combination of disinformation and deepfakes that can fool even the experts.

AUDIO RECORDING ALLEGEDLY BELONGING TO BIDEN

The latest examples of AI-generated election-themed content, including a series of images and videos circulated before the elections in Argentina and a manipulated audio recording of US President Joe Biden, are likely harbingers of what could happen on a larger scale.

WORLD ECONOMIC FORUM HIGHLIGHTS THE RISK

Misinformation and disinformation have been identified by the World Economic Forum (WEF) as the most significant global risk for the coming period. The problem with deepfakes is defined by the fact that AI-powered technology has now become cheap, accessible, and powerful enough to cause damage on a large scale.

This increases the ability of cybercriminals, malicious political actors, and hacktivists to launch persuasive disinformation campaigns and more transient, one-off scams.

According to the WEF report, synthetic content will manipulate individuals, damage economies, and fragment societies in various ways over the next two years.

The fact that tools such as ChatGPT and easily accessible generative AI (GenAI) enable a wider range of individuals to participate in the creation of disinformation campaigns driven by deepfake technology poses a significant challenge.

COULD BE A GAIN FOR RIVAL GROUPS

During an election process, deepfakes can be used to undermine voters' trust in a particular candidate. If the supporters of a political party or candidate can be conveniently steered by fake audio or videos, this would be a gain for rival groups.

At the heart of this challenge lies a simple truth: humans tend to value quantity and ease of understanding when processing information. The more content we see with a similar message, and the easier it is to understand, the higher the chance we will believe it.

When you add to this the fact that it is becoming increasingly difficult to distinguish deepfakes from real content, you have a recipe for a potential democratic disaster.

WHAT ARE TECHNOLOGY FIRMS DOING ABOUT THIS?

Both YouTube and Facebook are said to have been slow to respond to some deepfakes aimed at influencing a recent election. This occurred despite a new EU law (Digital Services Act) that requires social media companies to prevent attempts at election manipulation.

OpenAI, meanwhile, has announced that it will implement the Coalition for Content Provenance and Authenticity's (C2PA) digital credentials for images generated by DALL-E 3.

Cryptographic watermarking technology, also being tested by Meta and Google, is designed to make it harder to produce fake images. However, these are still baby steps, and there are valid concerns that as election fever grips the world, technological intervention against the threat will be too little and too late.

It will be difficult to quickly track and debunk fake audio or videos, especially when they spread in relatively closed networks such as WhatsApp groups or via robocalls.