Artificial Intelligence (AI) has wisely weaseled itself into our day-to-day lives. It was inevitable that it would penetrate the fabric of society at some point. It is no wonder that discourse around this “snake oil” has grown. Last month, I was speaking at a panel on a session titled “Artificial Intelligence and the Integrity of African Elections” on the sidelines of the Data Fest Africa 2024 in Nairobi Kenya.
We explored ways AI can enable and impact the democratic processes in Africa and with over 17 countries on the continent gone or going to the polls this year, I wouldn’t think of a more relevant time to dissect what this technology means for African democracy.
The good, and the not-so-good of AI in elections
AI, of course, plays a key role in the administration of elections with functions ranging from how we do voter engagement, ensuring transparency etcetera. This has the potential to enhance electoral processes. But this comes at the backdrop of some arguments that AI is merely a tool and nothing more. On the flip side, this double-edged technology offers rogue actors avenues for misuse and can threaten the very foundation of democracy.
For example, we have seen a rise in AI-generated deep fakes, which is a portmanteau for “deep learning” and “fake” entirely fabricated media, representing one of the most perfidious ways electoral sanctity is under threat/attack. These manipulated media forms can be used to confuse and dissuade the public regarding the position of the candidates, attribute statements, and even create scenarios that never occurred—all undermining the ability of the voters to make informed decisions.
The dangers posed by deepfakes are apparent. In the last few years, we’ve witnessed AI-generated content being weaponized in political campaigns in Africa and globally. Just this year, we saw it at play in the Indian general election with a fake video of M Karunanidhi, in which he was cheering on his old buddy TR Baalu at a book launch. Yet Karunanidhi passed away in 2018. This video, though debunked, had already sown seeds of doubt and confusion among voters, highlighting how AI can be used to manipulate public perception and disrupt the democratic process.
Where we have come from …
It is important to understand where we have come from as a people that use and interact with AI as a technology and the challenge it poses for the election process and democracy at large. Like I’ve said before, this whole AI thing has completely changed the game. In the past, we have seen smaller campaigns of misinformation and disinformation that were less sophisticated.
For those who have not watched the motion picture “The Great Hack” (Netflix), the scale of social engineering in the 2016 US Elections was merely a peek in terms of a certain scale into the era being ushered in. In comparison to what we have today, that was a simpler manipulation campaign because, of course, today we have a new variable that is AI which heightens manipulation.
Currently, AI-driven disinformation is complex. The opportunity to transition from traditional mis/disinformation campaigns to modern ones enabled by AI is a worryingly interesting area to explore. The power of sophistication and speed in terms of both generation and dissemination of skewed narratives has reached a worrying level.
Recently, the volumes of AI-generated content have been an issue since this overwhelms information ecosystems which overloads and makes it very difficult for the unsuspecting public to sieve through the real and fake of all the content rendering them susceptible to these false information campaigns.
The current state of interplay
There is a promise that AI will revolutionize how we carry out the administration of elections in terms of ballot verification, information dissemination, and whatnot. But all this is hype given that African governments have tons of priorities and free and fair elections are not one of those priorities.
So of course, there will be little or no budgetary allocations for these sorts of endeavors, and yet resources are a key enabler for how we as a continent (Africa) conduct processes that usher in new governments. The other issue is around the weakness of our technological infrastructure as a continent that leaves it susceptible to foreign influences and further complicates efforts to ensure fair and transparent elections.
There is a promise that AI will revolutionise how we carryout administration of elections in terms of ballot verification, information dissemination and what not.
The unchecked use of AI in the election cycles has resulted in voter manipulation. But unrelatedly, yet of concern, is the power that AI affords rogue cyber actors to cripple the information and electoral infrastructure of a given country. For example, with AI, it would be easier given the power and faster given the processing speed of these technologies for state-backed/sponsored rogue as well as independent cyber actors to attack and bring down the electoral infrastructure in an election period in that country.
We have seen increased polarisation of the electorate from simple AI-enabled content shared on social media platforms. Imagine the impact this will have on society if these technologies remain unchecked in election seasons with the increased tensions between rival factions in a race. This is why we need to swing into action and ensure the protection of the sanctity of the election season.
However, this discussion doesn’t negate other intervening variables, that affect the integrity of any election here in Africa and in all parts of the world. We have hacks from domestic and foreign agents, issues around the apparent digital divide, education, human rights abuses during the election cycle etcetera.
This discussion doesn’t negate other intervening variables, that affect the integrity of any election here in Africa and in all parts of the world.
Way forward
We need to leverage the opportunity presented through policy and regulation. Legislation has the power to influence societal conduct. We can move for laws that regulate the use of AI in election cycles to avoid and bar voter manipulation. In the United States, this has been successful where some states have moved for disclosure when AI is used in the generation of campaign content or some states, a total ban on the use of AI in the election cycles.
I am a strong advocate for the use of AI with guardrails in place to preserve democracy, residually —human rights, the rule of law, and society. We should be reminded that regulation is always determined by the incentives of those behind it. Even then, I do not think it is wise to entirely ban the use of AI in these sorts of processes.
See also: OP-ED: Scared of AI? It Won’t Replace Humans
This helps protect the integrity of the election cycle. I am a strong believer that everything flows from thorough policies and also placing Africa as an ideological leader as compared to previously where we have applied “copy and paste” of laws and models from foreign jurisdiction. It would be a great opportunity for us to set the agenda and shape the narrative around the regulation of AI use in election processes.
Education and awareness by CSOs, publishing, and information powerhouses like PC Tech Magazine to supplant the efforts of the electoral commission, are very vital in ensuring that we have an educated and informed public. An informed public is an empowered public and will be in place to spot and debunk some of these mis/disinformation campaigns by rogue domestic actors, nations’ autocratic regimes, and other foreign actors who might want to interfere in the electoral processes of a nation.
But again, political will is also necessary for both the legislative and digital literacy agenda to take off given that the process of law-making, implementation, and education starts with adequate resource allocation which requires a vote at the budgetary level. Given the disparity in national priorities, I think it would be wise for governments to rethink and find a healthy balance between conflicting priorities.
In the spirit of unity, cooperation between governments, civil society, and the private sector is crucial. This is simply because the process of enacting standards and codes to govern the use and or guide the extent of and application of artificial intelligence requires concerted conspiracies between the government, CSOs, academia, and the private sector. International cooperation efforts are important since the challenges posed by AI can be transboundary at times.
In particular, making a disclosure, transparency, and tagging certain information as AI-generated or deep fake may help in the taming of this scourge. This goes with ensuring responsibility by design from the developers’ end when these AI models are developed. It ensures that they inherently abhor misinformative and disinforming rhetoric and campaign gimmicks that jeopardize democracies. This has been very successful in several states in the United States and a few other foreign jurisdictions where it has been piloted.
Civic space actors, organizations, and others can leverage open-source intelligence in trend analysis, fact-checking, and probing data/information that makes rounds during elections. Fortunately, millions of OSINT resources are available on social media and are easier to use compared to the traditional fact-checking and verifying avenues.
Organizations like the Citizen Report are doing commendable work in terms of fact-checking information that is disseminated on social media and elsewhere. Such initiatives can supplement OSINT technologies. OSINT has also been a great aid in differentiating videos produced by the propaganda machinery of Israel/Palestine or Russia/Ukraine in their offensive. A true demonstration of its ability to cure mis/disinformation shenanigans.
Lastly …
AI in electoral processes is bound to make democracy stronger or to erode if left unchecked. How well we navigate will dictate the future of African democracy and back home, the impending 2026 Ugandan elections. Let us focus on strengthening legal regimes, bettering the degree of digital literacy, as well as fostering cooperation that would help secure the integrity of elections and protect human rights, society, and democracies. This should, therefore, serve as a summon to all concerned actors to ensure the technology is directed towards working for a stronger democracy in Africa.
