Artificial Consensus
- Campaign On Digital Ethics

- 12 minutes ago
- 6 min read

By Kristen Abrahams
I am an intensely competitive person. Whether it is a game of padel, a school test, or (most importantly to those who know me) a round of Thirty Seconds, I have always loved winning. This competitiveness extends beyond games. What irks my family most is my conversational persistence (they call it stubbornness; I call it spirited debate), because I simply cannot let an argument go. If you can relate, you will understand why nothing has empowered me quite like the internet. There is a particular satisfaction in saying ‘just Google it,’ proving your point, and ending the conversation there.
Generative AI has taken this dynamic to another level entirely. When even Google results are disputed, few people argue once ChatGPT has weighed in. This method of determining fact from fiction has begun to shape our interactions with one another in subtle but profound ways. ‘Ask Chat’ has become a commonplace refrain at our dinner tables, in our workplaces, and even in our most intimate conversations. We have all, to varying degrees, come to trust artificial intelligence. But have we paused to consider the long-term consequences of everyone having access to the same information? A relatively new series called Pluribus probes this question.
The Hive Mind
Without giving too much away, the series imagines a dystopian future in which a genetic sequence causes everyone, except a rare few, to think with one mind. If you are a reader of the Bible, you may recognise this kind of language: it recurs throughout the text, encouraging believers to be unified and to reject division among themselves. Pluribus, however, depicts a more unsettling interpretation. In the world it portrays, everyone has access to the same thinking, the same knowledge, and the same values and beliefs.
This manifests in striking ways. An eight-year-old can perform heart surgery because he shares a mind with the world’s leading cardiovascular surgeons. He would equally know what V.S. Naipaul was thinking when he wrote A House for Mr. Biswas, because he shares Naipaul’s mind too.
There is also a shared moral consciousness. None of ‘The Joined’, as the affected humans are called, kill for food or pick an apple before it has fallen from a tree. Everyone loves animals, tends the environment, and is invested in everyone else’s wellbeing. These commendable moral standards notwithstanding, Carol, the show’s protagonist and one of the few not yet joined, refuses to undergo the procedure that would merge her mind with ‘The Hive.’ Drawing on her harrowing experience of conversion camps, Carol invites us to examine our own methods of ‘converting’ others to our brand of politics, our ideals, or our moral beliefs.
The Danger of Consensus
It is well established that the data fed into generative AI systems shapes what those systems produce. When X began amplifying disinformation about a so-called ‘White genocide’ in South Africa, we saw how readily millions of people accept something as true when generative AI appears to confirm it. This is obviously dangerous when those systems are, by design, racist or misogynistic. But what if they were trained exclusively on progressive ideals?
The world of Pluribus explores exactly this tension: a society of complete harmony, achieved at the cost of individual autonomy. When Carol is offered this harmony at the price of her individuality, she refuses and implores the other immune humans to do the same:
“Now, I realise some of you think the world might be better off this way, with all the newfound peace, love and understanding. Enjoy that opinion! Relish it! Because it may be the last one you ever possess. And when the day comes that you have peace and love forced upon you, who knows, maybe in that last fleeting moment… you might just realise you treasured your individuality.”
Would those of us who consider ourselves progressive be content to convert the entire world to our brand of thinking, even if it meant surrendering individuality, if we were guaranteed a constant state of peace and unity? These are the questions we must confront as we navigate a world in which the large language models powering our AI tools can, quite literally, nudge all of us towards the same thoughts. That is obviously dangerous when a model is trained on misinformation or hatred. But is it not equally dangerous to present progressive politics as uncontested fact?
The Virtue of Grappling
This piece is difficult to write, particularly given the state of the world we are living in, and for many, merely surviving. With the rise of populist nationalism, fundamentalist religiosity, and naked hatred, it seems almost naïve to suggest we should engage with those with whom we fundamentally disagree – with those who spread hate and vitriol, who are responsible, directly or implicitly, for not only threatened harm but the active destruction of entire communities. This piece does not offer easy or definitive answers. It offers a perspective, one that may prove to be wrong, though we rarely know what is right or wrong in such stark terms until we see the results of our actions.
Forced tolerance is not the same thing as unity. All it means is that some people’s positions have frightened others into adopting those positions. Does that sound familiar? Or does it not matter, so long as those positions are in line with our own views? Are we not called to engage with the person behind the idea, to understand how and why their views were formed, and to seek understanding before we judge – before we decide that a person is unworthy of engagement?
If we ask a chatbot for advice and it teaches us to accept opinion, good or bad, as fact, rather than pushing us to engage with the human being in front of us, is that not the real danger it presents? Regardless of whether it is feeding us progressive insights, is the greater value not found in arguing, grappling, and learning from others? I remember grappling with ideas I now consider fundamental. I was given the space to struggle, to feel, and to be wrong, because the people engaging with me were not judging me. They were getting to know me.
Democracy as an Ongoing Grapple
I was born in 1994, the year apartheid formally ended and South Africa’s democracy began. What that transition required, among other things, was for people on opposite ends of the spectrum to sit at the same table, listen to one another, and learn from each other. To set aside deeply ingrained biases and very real suspicions, and to truly engage with the person in front of them. Whether this was done on both sides remains open to debate, and we should remain critical of the positions parties took and the costly concessions that were made.
There is, however, something to be learnt from that process. Those who hold different opinions, worldviews, and beliefs from us are still human beings: their views are learnt, not innately inherited. In a country like ours, engaging with people who think differently is not only necessary but inevitable. We interact with people from all walks of life every single day. We need them to interview us, to treat us medically, or to hand us our change. And when we listen to their perspectives, truly and openly, we may find broader and more positive results than we anticipated.
Our South African democracy is only as young as I am. And while it is easy, and often justified, to deplore the state of our country, we must also pause for reflection. We have all navigated this journey through grappling: with disappointment in our own party, with sometimes misplaced hope in others, with breakaway parties, coalition governments, and fresh faces on the political landscape. Yet isn’t this ever-shifting landscape, marked by a plurality of views and diverse constituencies, not something to be celebrated rather than resented? Is there not more to be gained in understanding why someone votes for Tannie Helen than in dismissing their choice as ‘unwoke’? Does it not mean that our democracy, in holding space for the good, the bad, and the outright ridiculous, is – dare I say it – working?
As we enter the local government election season, in our interactions with political candidates, first-time voters, and long-standing activists, may we remember that all of us had to grapple with a new way of being after 1994. That as we navigate this complex world and its people, we are continually grappling and learning. That everyone has something to teach, and that we all have something to learn. May we find greater value in human interaction – with all humans – than in hard and fast, definitive answers. May we stay curious and critical, but always human. And may we afford others the grace that our democracy and world require, as we expect others to do the same for us. Otherwise, although we all may end up thinking with one mind, is it worth it if we all end up devouring each other to get there.[1]
Kristen is a Legal and Policy Lead at the Campaign On Digital Ethics | Kristen@code-sa.org
[1] Spoiler alert: as food runs out and resources are scarce – because cherries can’t be picked before they fall and sheep can’t be slaughtered – Carol begins to investigate how ‘The Joined’ are surviving. In a shockingly revelatory episode, Carol discovers that they have been turning the remains of deceased humans into nutrition, in a powder-like form (think collagen, but human).



Comments