What it takes to fight increasingly sophisticated, AI-powered disinformation campaigns

Decades ago, a desire to report on South Africa’s journey pushed me to become a journalist. I joined the BBC believing people should have access to facts they can trust. Facts that would help them make important life decisions. Facts that would nourish democracies around the globe.   

I never imagined we’d land up here.  

Iran stands accused of building a “stable of fake journalists” to spread lies online. OpenAI says a Russian influence operation recruited unwitting researchers in Latin America to run a front think tank. Britain is planning to open a new national centre to fight disinformation from hostile states, which Prime Minister Andy Burnham says “reaches into people’s homes and twists what they feel … stoking division and sowing despair”.   

A report released earlier this month shows us how, in Latvia, a real military exercise was used as a seed of truth in a malicious fake video of a shooting that never took place. The video falsely claimed that a Ukrainian soldier had shot and killed three Nato colleagues before taking his own life. According to the report, there were at least eight different versions of the video in 15 languages, sharpening its ability to cut through to audiences around the world. AI was used to mutate the original video.  

Bot farms in Africa are amplifying harmful content, and AI models can now manage social media accounts. The real-world damage is mounting, elections are being influenced and important institutions are being undermined. South Africans need no lesson on where the road through broken oversight institutions leads. Nor on how important free, independent and trusted journalism is to a democracy.  

When I launched the BBC’s Trusted News Initiative, we brought together big tech and global media, including Daily Maverick as a founding partner. We were creating filters and alarm bells for potentially harmful content. That curation feels quaint now. What we’re seeing is full-blown information warfare, supercharged by AI.  

It’s been more than 30 years since my producing assignment in South Africa, covering Nelson Mandela addressing jubilant crowds in an election that finally consigned apartheid to history. The technology around us has evolved, but my belief in facts-based information hasn’t changed. That’s why I founded Athena, a publication that maps out the terrain in the fast-moving information war. By speaking to experts on the frontlines of the information war, we’re starting to piece together solutions – ways for nations, companies and individuals to reinforce their armour against disinformation.  

How do we do that? 

In the UK, we speak of whole-of-society resilience against attempts to manipulate us. Norway, sharing a border with Russia, has long had a sense that truly effective defence starts with ensuring that everyone in the country is aware of the threats and what their part in countering them should be. Defence in the age of hybrid warfare (physical, cyber and information attacks) has to be a national effort. Ukraine learnt that very quickly.  

Three ways forward

Whole-of-society resilience is a huge topic. But let me put forward three possible ways of bringing the whole society into this conversation: 

Speak about the unspeakable. That was a great line from Deborah Haynes, defence and security editor at Sky News. She was talking about Britain discussing the “w-word”: war. Her point was about a national conversation taking place but, she said, to make that happen means a government that has the confidence to share information with the media and the public. Haynes called for more intelligence to be made available to drive out malevolent disinformation. An effective press must be free to scrutinise intelligence – and government in general – including when that scrutiny is uncomfortable. 

Disinformation wargames. Critical infrastructure (power, water, banking, airports) is run by public and private bodies. Far more work needs to be done linking those operators with government authorities at local and national level. Together, they need to wargame different attack scenarios, and tell the public about it. 

Linked to this, companies must prepare. They are vulnerable and need to think urgently about everything from coordinated disinformation campaigns that seek to undermine their brands to attempts by bad actors (be they state-sponsored or criminal) to pollute the data of companies with false information. Cybersecurity has become a responsibility across organisations. Disinformation now demands the same level of awareness. 

Inoculate the nation. What should public awareness of disinformation look like? Everybody needs to be able to recognise when they are being manipulated (remember the divisive campaign exposed in the GuptaLeaks) and how social media is being flooded with so-called inauthentic material.

AI is so good now that people can no longer rely on identifying deepfake videos or audio simply because they look or sound odd. They need to be given the tools to work out for themselves why someone might be trying to manipulate them, how they might be doing it, and what they can do about it. We have lots of ways of providing those tools, but it will mean harnessing the power of bodies that can reach deep into our society: journalism, yes, but also organisations like schools, local libraries and community centres.  

Information warfare may feel distant. It may not feel like a priority when there are so many other problems: unemployment, corruption, inequality. But insidious disinformation campaigns distort and divide nations, paralysing desperately needed progress. Solving these problems takes participation across society, preparation and reliable information. DM 

Follow Athena on Substack.