Deepfakes: journalism, media and democracy in the age of AI

Ylva Rodny-Gumede is the head of the Division for Global Engagement and a professor in the School of Communication at the University of Johannesburg.

She recently published an opinion article that first appeared in the Mail & Guardian on 22 November 2024.

This week a titillating video went viral on social media featuring former health ombud Professor Malegapuru Makgoba claiming to have found a cure for diabetes. Absolutely fantastic — if only it was true.

The video was altered with artificial intelligence (AI) and while real and credible videos exist featuring Makgoba, he has never spoken about diabetes in any interview, let alone in one where he asserted that he had found a cure for the disease.

Deepfakes and synthetic material — material that resembles reality or real people but is generated artificially — is becoming more prevalent on social media and sometimes also makes it into the mainstream news flow.

Fake news is increasingly shaping the way in which public discourse is formed about politics, elections and a range of social and health issues.

Conspiracy theories, and even incitement to violence, are being peddled as news and the truth. Sometimes the gain is purely commercial, to drive traction and audiences to a particular platform, and sometimes it is to drum up support for a political party or cause.

Either way, it is cynical and always comes with a huge cost to truth and democratic deliberation in the interest of peace and prosperity. In short, however we label it, fake news threatens the very basis on which modern democracies are founded and upheld.

Of course, none of this is the fault of “AI” — it is how we have let algorithms steer processes and data collection and the dissemination thereof. But generative AI does take control of programming out of our hands because it centres on learning, innovation and the ability to create and synthesise rather than analyse and predict.

AI is altering societies fundamentally, and no more so than in how we interact with each other and understand ourselves, as well as state formation, and our own roles and places in society.

The link between media and democracy has often been questioned, particularly considering the problems facing traditional media. Traditional media was long thought to be the main conduit for information between the public, governance and governing parties in society, and the so-called public sphere often thought of as the main pillar and forum for democratic deliberation.

The questions are: what information flows are keeping contemporary democratic debates and deliberation alive and what is the role of new media outlets, social media platforms and AI in shaping democracy? And also, what happens when journalists, as conveyors and intermediaries of news, as well as professional and authoritarian voices, are discredited or even annihilated?

The lack of transparency in AI systems raises major concerns about biases and errors creeping into news stories — and this without deepfakes or intentional spreading of disinformation. Fact-checking remains key and generative AI does not differ from any other information source, human or other, in that it has the capacity to mislead.

AI algorithms that have generated fake news, and even spread messages of hatred and incitement to violence, have been triggered by social media platforms seeking to gain audiences based on “news” feeds that gather traction. Thus, the old adage “bad news sells” still holds sway.

And although AI can be seen as a tool that allows journalists to do their research better, this might be offset by additional demands on them in a resource-starved environment that mitigates oversight and fact-checking of AI-generated data.

In addition, considerable investment is needed in training journalists to use new technologies, and generative AI in particular, as the speed of development — as creative and adaptable as we are as a species — outruns us and our ability to comprehend and foresee what lies ahead.

On the upshot, for journalism AI’s potential to increase efficiency in news organisations is crucial but might still not be all that it is thought to be. AI can definitely assist with editing some research, particularly as pertains to big data analysis.

For the news business model AI can support efficiency and productivity and dynamic paywalls are crucial for financial sustainability.

What needs to happen is for news media outlets to take control of developments and to partner with tech companies that provide both the AI platforms and tools for the detection of AI-generated content.

There is also a real need for broader societal conversations around fake news and our responsibilities as citizens to guard against spreading false news and misinformation.

Media literacy has never been more needed and understanding how to use social media responsibly needs to become part of the curriculum from primary school through to university.  And we need to move quickly.

Generative AI, left on its own, could wreak major havoc in the way that we as societies impart and receive information and how we communicate as individuals, communities, nations and as a global community made up of sentient beings.

We need to ensure that communication is between ourselves and not algorithmic agents. Algorithms are a tool, and only a tool, it is up to us to find a way to harness the good and regulate the bad in the interest of free, open and meaningful democratic deliberation.

*The views expressed in this article are that of the author/s and do not necessarily reflect that of the University of Johannesburg.

Share this