We are now two short years before the 2027 elections. Like all of Kenya’s elections where an incumbent is seeking re-election, the 2027 contest promises more than the usual antagonisms and hostilities.
Going by the reckless statements already being made about traitors and betrayers, there could even be violence, mirroring events of 1997, 2007 and 2017. Even as the state and non-state actors consider how to reduce the possibility of election-related instability arising from these realities, there is yet another phenomenon which I believe requires attention.
This phenomenon can grossly infect the elections and be the catalyst for even more hostilities. That is the role of Artificial Intelligence-generated propaganda and its capacity to undermine the democratic content of the election and inject content that is destabilising.
A recent study entitled “AI propaganda threatens African democracy” outlines at length the dangers posed by Artificial Intelligence on Africa’s democracy.
It shows how AI-generated content negatively impacted African elections including those in South Africa, Nigeria, Senegal and Gabon. This challenge is particularly acute for a country like Kenya that is a global leader in internet access and social media engagement.
Many citizens depend on the internet as their source of information and news. In the last five years, generative AI has become more sophisticated and is now the source of information and news for these internet users. Its use in political activity has risen to unimaginable levels.
But for all its good uses, it has also been applied to undermine elections inter-alia through deep fakes that are not easily detectable and therefore grossly misleading. The 2023 elections in Nigeria were the first in Africa where widespread use of AI propaganda shifted opinions and infected the elections. In the last weeks of the elections numerous “endorsements” of political figures from international celebrities, local business leaders were common. Reports of “newly discovered scandals” about politicians were legion.
The worst of the propaganda, which I expect to emerge in Kenya’s elections, were “exclusive exposes”; fake videos of candidates purportedly planning to rig the elections. It reminds one of the hysteria in 2007 when various places were raided and burnt down purportedly because rigging was being undertaken therein. In the Kriegler report of 2008, it transpired that these were well planned acts that the instigators knew were fake.
Imagine if such allegations had been supported by deep fake videos showing rigging plans by significant political players. Unfortunately for all the widespread use of the internet globally, users are still largely unsophisticated in identifying fake AI content. I have watched with amazement typical stories liberally circulated which are obviously fake but even the more sophisticated internet users assume to be true.
The most common relate to the new generals in West Africa particularly Ibrahim Traore, the Burkinabe strongman. Numerous fakes exist of him making all manner of “progressive” anti-West pronouncements and leading his country into Nirvana. These videos have led to a surge in support for military strongmen. Research on these and other videos that target Africa show that they emanate from Russia and China and are part of the new cold war supremacy wars pitting these countries against the West. How is this threat to be countered?
The reality is that it is almost impossible to totally weed out fake material from the internet. It is expensive and time-consuming work which explains why companies like Meta have reduced their investment in fact checking. In any case AI is increasing in volume and getting better at concealment making reliable fact checking impossible. While countries must continue to invest in fact checker programmes, the ultimate answer is civic awareness. Governments must invest in targeted public education on identifying AI generated content. A more alert public will reduce the impact of AI propaganda.
More critically there is need for a continent-wide approach to counter this external threat to Africa’s still fragile democracy by limiting its entry into the continent. This approach must emanate from a recognition by African governments that generative AI, for all its progressive potential in various sectors, has capacity to undermine public trust, democratic processes and ultimately lead to political stability, the last thing an already struggling continent needs.
-The writer is an advocate of the High Court of Kenya
