Hello there, dear reader of mine, I have a confession to make.
I, Monique Kwachou, writer, holder of a doctorate, educational counsellor, and all the things you know me for, use Artificial Intelligence, AKA AI.
I feel like a lot of us don’t want to confess that. I mean, it is not exactly a secret. We all know we use it, but it seems we are all supposed to pretend we don’t. And when we are upfront that we use it, we tend to be vague about how exactly, and how much. Why? Because it raises questions about our intelligence and the originality of our work, and it casts doubt on our ability as a whole. Suddenly, every single thing you have written is scrutinised. Is it AI, or is it you? How much of your thinking is actually yours?
I have been musing on the complexity of being wary of AI dependency and cognizant of the risks it poses, while also understanding its usefulness to our generation and the fact that it is not as new as we make it seem. And as with most of my musings, I thought to put this out there for public engagement.
First, I must admit that I am writing this from a position of some privilege. Have you come across the tweet that says the millennial African parent’s version of “Oh, I used to walk 5 kilometres to school and back. You people are so lazy, you don’t know how easy you have it” is going to be us telling our children, “I used to research without AI. You kids don’t know how easy you have it”? Well, as a millennial, I can even say I used to research without Google. Because millennials, we have a flex, okay? We did libraries. I know how to use a library. I know the order in which the books are placed on the shelves, because we actually had to find them ourselves. We came of age with the internet. We used Ask Jeeves before there was Google, then Google came and we learned to Google things, all the way to the point that we are now using AI.
I think that is partly why I have some confidence about where my own thinking ends and where the assistance begins, and the audacity to put these musings out there. Having earned my PhD before generative AI was a thing, there is little room for anyone to question whether I’m smart, myself included. So I feel comparatively safe making this confession and discussing what I think is being overlooked in conversations about AI use and dependency. And yet, even with that relative security in my own ability, that niggling voice of interrogation still lingers. Are you using AI too much? Are you using AI ethically? Should you have done this without any assistance? And the guilt around it.|
So I needed to address it and be honest about my own use of AI. But I also wanted to add something to this whole AI conversation that I don’t think we are talking about enough.
Something about the way we are talking about AI right now keeps reminding me of what I see all the time in development work and social justice work. People complain about the fruit of the issue, the leaves, what we can see at the top of the tree, but they do not actually want to address the root. And if you don’t want to ask why the thing is happening in the first place, I don’t think you really want to solve the problem. That’s what I generally believe, anyway.
With AI, the leaves are everywhere. There are countless books being published with fake authors, AI-rendered flyers are like eyesores across our feeds, and both our kids and our elderly are taking AI videos and answers as fact. And don’t forget the ever-present environmental threat of the data centres that keep these things running. I get it. The widespread use of AI is concerning, and these problems are real and valid to criticise. That fact is not to be followed by a “but,” rather an “and.” AI poses countless problems, and still, the way discussions about its use are going feels puritanical and hypocritical. I’ve heard this kind of judgement before, from those who say people who have had bariatric surgery or use GLP-1s are on a “fake fitness journey,” or who accuse women who have C-sections of taking “the easy way out.”
Those judgements not only ignore what would lead to such decisions, but also suggest there is some sort of award for taking the hard way even when assistance is available… like how we African women are often told to prove our worth by how much we endure or how much effort we have made to earn value, this rubs me wrong. And that is my problem with the AI conversation. I am seeing so much discussion about what AI is doing to us, and not nearly enough about why so many of us are reaching for it. I feel like we are blaming individuals for something that is systemic, and we know it. If we keep looking only at the leaves, we fail to see what is causing this dependence in the first place.
So here is what I want to submit: the increasing AI usage and dependency we are so worried about has roots in the kind of world we have built and those roots must be examined. I have considered a few interconnected roots and invite you to consider them along with me.
First, let’s be honest, this is a natural progression. The outsourcing our thinking is not new.
People love to say, “We all have the same 24 hours.” We do not, and I think to even suggest that is gaslighting.
I have somebody who comes and helps me clean my house twice a week. I am also a single woman without children. I do not have the same 24 hours as a working mother of three. Her 24 hours are considerably more heavy-laden than mine, and she might barely have enough time to think. So if you tell that woman, here is a virtual assistant that can help her draft up a meal plan, do household budgeting, or make her workload a bit lighter so she can clear her desk faster and spend more time with her family, of course she would be grateful for that.
And money and domestic responsibilities aren’t the only things that change how much of those 24 hours we actually have. As somebody with a hearing impairment, my undergraduate experience at the University of Buea (where the average classroom size was over a hundred students) meant I had little accommodation, as lecturers dictated notes from the front of amphitheatres. I couldn’t listen and take down notes at the same time. So I would listen, watch the teacher, and concentrate, then go home and make my own notes, borrowing books from classmates and doing additional research to verify what they had written. In some ways that probably helped my studies, because I had to go over all the material twice. But that is exactly the point. I was attending class twice. I did not have the same 24 hours as the student who could hear the lecturer, take notes and leave the classroom with the work essentially done.
Why are the above illustrations relevant? Well, because the myth of us having ‘the same 24 hours in a day’ hides invisible labour and overlooked privileges very well, leading to another false assumption: that outsourcing intellectual labour is something new, something that just came up with ChatGPT. It is not. We know for a fact that historically great men have had their wives doing a whole lot of the planning, the support, and yes, the thinking, sometimes even helping with the writing itself, while they took the credit. Even where their wives’ contributions were limited to taking care of the home, the children, taking care of life, that outsourcing of thinking about what to make for dinner made it so they had time to sit somewhere and think and write and invent things.
Likewise, academics have had research assistants forever. I mean, some of us were the research assistants! Professors have long relied on underpaid, undervalued research assistants and doctoral candidates, going to the field, doing references, transcribing, collecting data, and too often that work has quietly become the professor’s work. Senior stylists and designers have done the same with their junior colleagues, senior architects with their junior architects. Lawyers have had legal aides doing their research for them. And as a scholarship recipient I know for a fact that we have had a stipend made available for the sole purpose of paying somebody to edit their PhD before submission, like most authors do their book before it goes to the publisher.
If we are honest, the pattern of leaning on someone’s undervalued labour hasn’t changed. The tools we are all now using were trained and cleaned up by low-paid data workers, many of them in the Global South. There has been reporting on workers in Kenya paid less than two dollars an hour to sift through some of the most disturbing content on the internet so that the rest of us could have a cleaner chatbot.
What has changed is the who is being undervalued and the who gets to benefit from it. We have moved from the wife, the research assistant, the junior architect, the legal aide, to now the data worker in Nairobi. Likewise, the ability to and practice of outsourcing has simply become more affordable and accessible to the average person. And I feel like the backlash we are seeing is, at least in part, about exactly that.
It reminds me of industrialisation. When machines came along that made money for men and for the upper class, nobody had a problem. But when machines started making domestic labour easier for women, the washing machine, the microwave, suddenly it became an issue of women becoming lazy. In all of this, we are not questioning the system. We are questioning the person who deserves help, and not the fact that help has always been needed.
We’re uncomfortable with the fact that the layperson is getting the help that was only available to a few, seemingly more ‘deserved’ persons who could afford it. That is the danger of Large Language Models (LLMs) being made for the mass market.
Which brings me to the next root issue…
Who actually gets questioned for using assistance? Who gets suspected, who gets a pass, and why?
If the problem were simply AI, you would expect the suspicion to fall evenly. It doesn’t.
Think about who is most often asked, “Is this AI, or is it you?” The AI detectors that schools and universities are relying on to catch AI-plagiarised work have been shown to flag writing by non-native English speakers as machine-generated far more often than writing by native speakers. So the people most likely to be accused are the African scholar writing in her second or third language, the student who never had a paid editor, the very people that AI tools such as Grammarly are offering to level the playing field for. To be fair, we have even reached a point where books published centuries ago can be accused of AI plagiarism by the wrong detection tool, which is so funny to me and yet horrendous at the same time. If a writer from centuries past can’t pass the test, what chance does a Black breakout author have?
This particular root of the AI dependency problem is embedded in the history of colonialism, and in the continued, pervasive coloniality that formerly colonised people are constantly being told to get over. Whether we like it or not, we have said there is a particular way we should have knowledge, and a particular way knowledge should be presented. This is the rubric for academic writing. This is the rubric for good English. This is how you should sound. This is how you should present your argument. So if we are seeing people turning to LLMs to write better, to generate knowledge in a particular kind of way, or simply to sound smart, it is because they are trying to meet that rubric. And that is also why these LLMs present arguments in the same voice and the same tone, so much so that we can even catch it.
To be honest, we don’t really want variety. The rollback of all diversity, equity, and inclusion programs is evidence of the fact that we don’t really want that. We want everybody to be the same, to think the same, to be compared with the same rubric. We have raised people to know that uniformity is rewarded. Fitting into the colonial matrix is rewarded. So you see that the AI flyers all look the same, and you see that people increasingly sound the same.
If creativity and different ways of knowing were genuinely encouraged, if we were actually decolonised, I don’t think we would be depending on these tools so much.Or at least we would be seeing more original and varied use of them. Even the flyers would be more creative.
And the true irony is that many of the institutions condemning those who use these tools to meet the standard are the very same institutions that set the rubric and are embracing the technology themselves. Universities that penalise students for using AI are backing data centres and rolling it out themselves. Employers who frown on staff using it are using it to cut jobs, speed up output and increase their profits. So the individual is policed for using the machine, while the institution is praised for innovating with it. The system that demands machine-like productivity from us is condemning the use of the machines.
And that brings me to the mother root of it all…
We built a world a world for machines, wasn’t the goal optimisation?
If we are being truly, truly honest, our society has long since decided we did not really want intelligence in itself. We want what intelligence could give us. Generating knowledge for its own sake? LOL! Even the research institutions no longer do that… We haven’t built a world where people are encouraged to think for themselves orbe creative. We have rewarded productivity. We have rewarded optimisation. We have created a world for machines and furthered individualism with that, so why are we shocked that people are depending on machines to live even more productively and individually?
And this did not begin with AI. Think about maps. We used to drive around with one, take the wrong road, get missing, stop and ask somebody, “Please, do you know where to find this place?” We used to ask the neighbour next door, the cousin down the road. But with end-stage capitalism, we have become so individualised, so used to self-reliance, that we now reach for whatever gets us there fastest, and Google Maps does. Or think about clocks. I see my nieces and nephew unable to read a traditional clock, that is likely to become a lost skill. But that is because we optimised to a digital clock that simply tells them the time in a way they can read. They don’t need the other one anymore. With optimisation, we made it clear the point isn’t the skill of telling time as our ancestors did from the sun, or as the currently ageing generation did with a traditional clock, butwhatever gets the answer fastest. Even if that means merely asking Siri or Bixby for the time.
All of this reminds me of a lesson I learned during my PhD. I had written my first analysis chapter draft and my supervisor basically told me no, this is nonsense, this isn’t going to work. And I was so frustrated that I couldn’t write. For close to three weeks I basically did nothing. I cried. I moped. I walked around. This was during COVID, so I was taking these long walks around campus, distracted, wallowing, not writing. And then one day, during one of those walks, the idea for what I needed to do with the analysis just came to me. Literally during the walk. I pulled out my phone and started recording myself talking it through as I walked using Otter AI, and when I went back to my room, I had the skeleton of the chapter and was able to write the new analysis chapter in less than a week.
I have never forgotten that, because it took me something like three weeks of seemingly doing nothing to get the brainstorm that allowed me to do the work in one week. Except I wasn’t doing nothing, was I? That was the work too. My brain needed to disengage. I needed to walk, to mope, to not produce anything for a while, for something to ferment enough for me to know what I wanted to say.
I remember talking to my co-supervisor and friends afterwards and remarking on this identified problem. You need time for inspiration to come, but we don’t have time, because we have deadlines. And those deadlines become evidence of whether you’re intelligent or not, hardworking or lazy, good at your job or not. But creativity doesn’t always work on deadlines; so the person who cannot wait three weeks for the idea to ferment, because the thing is due on Friday, asks AI to brainstorm it. And then we blame them.
We have praised multitasking. We have people doing five jobs by themselves, and we call it being hardworking. But you cannot endlessly multitask, because you are not a machine. So, of course, people outsource. Yes, there is a real risk that in outsourcing, we lose our thinking and our skills. But we optimise everything for better performance, so why are we shocked that everybody is optimising at their own level? We have built a world that demands machine-like productivity from human beings, and then we are horrified when human beings turn to machines to help them meet those demands.
And yes, there is a risk in all of this. We are losing our thinking, and we are losing skills. But why aren’t we considering that we are losing these things because everything now has to be optimised for efficiency? And optimisation is what you do when you are building a world for machines. We have created a system that expects us to operate as machines. We want speed, we want perfection, we want more, and we reward the person who produces more, faster, more perfectly. It has become the best robot against the best robot, Claude versus ChatGPT, and it is no longer really us. So why are we now upset that people are depending on machines? For me, that is disingenuous. We have built a world that demands machine-like productivity from human beings, and then we are horrified when human beings turn to machines to help them meet those demands.
Please Note: This is not a defence, but a call to deeper consideration and action
So far, this might sound like me making a case for AI or defending excessive use of it, in case anybody has reached this point and somehow that is their takeaway. Perish that thought! That is not the takeaway. I am trying to understand what I see as the root of why we have reached this point of excessive use.
Because there is still danger here. The person who genuinely does not have the time and the person who simply cannot be bothered now have access to the same tool. The mother of three trying to buy herself two hours and the student who does not want to read ten pages can both ask ChatGPT. And there are still so many things about AI that I hate. I hate that young people in isolation are asking AI for relationship advice and forming romantic connections with technology in ways that say something deeply worrying about how lonely we have become. And there are far darker uses of generative AI, involving sexualised images of children and other abusive material. So yes, this thing can be harmful. This is me saying, let us be honest about why we are here, so that we can address the problem better.
So what would it look like to address the root rather than the leaves? I don’t have all the answers, but I think it has to happen at three levels: the system, our institutions, and all of us at individual levels.
At the level of the system, if the problem is a world optimised for machine-like productivity, then part of the answer has to be making room for human beings again. That means more humane deadlines and workloads. Fighting the growing disrespect for the humanities and what is deemed ‘unproductive’. It means paying the people who do the assisting fairly, from domestic labour, to the underpaid graduate student assistant or the data worker in Nairobi. And it means holding these tech companies accountable for the water, the energy and the labour, instead of giving them the kind of leeway global leaders have given them so far.
At institutional level, it means assessment in our schools and universities that values the process of thinking, the drafts, the conversations, the reasoning out loud, and not just a polished final product that could have come from anywhere. It means protecting time to think, encouraging difference and uniqueness in ways of knowledge creation and expression, something the decolonising education movement was arguing for long before ChatGPT. Instead of policing individuals with detectors that are biased and unreliable, institutions could build a culture of transparency, where people can say how they used AI without fear.
At our individual level, I think it starts with stopping the shaming and starting the honesty. If we talk openly about how we use AI, younger people get to see what careful use looks like, instead of learning to hide it. And we need to rebuild the habit of asking people. Ask your cousin, your neighbour, your colleague, even when Google is faster. Read together. Mentor somebody. Take the long way sometimes just so you are confident that you can. We cannot individually undo end-stage capitalism, but we can refuse to let it make us completely alone.
And at my level, being honest upfront that I use AI as a writer and feminist scholar means I must constantly interrogate my use of it and note where I draw the line. My thoughts are constantly running: okay, Monique, don’t put highly sensitive personal information there, no uncoded research data, and so on. But beyond privacy, there are intellectual roadblocks I need too. I try to ask myself: do you ask AI first, or do you ask yourself first? Have you tried to write the paragraph, wrestle with the idea, and establish your own position before going to AI? I challenge myself to read as much as I use it. Honestly, I am failing at that challenge, but it is a commitment I am sticking to. I will fail and try again. I am more conscious than ever that I must read widely, and I play word games and other mind games to make sure my cognition is still exercised regularly. I never go to AI for personal advice either. There are things I should ask my people. There are things I should take to God. AI is not my Holy Spirit. Perhaps that sounds funny, but I have seen people turn AI into their best friend, closest confidant and conscience in worrying ways; obviously the Pope has too given his warning about use of AI for sermons. Never forget that you are the one giving it the commands, Monique, and never forget that you could do without it. Even when it is editing you, remember how you sound. How do you sound, Monique? Make sure you still know.
For me, ethical use also means rejecting unnecessary AI use. AI music? I hate those songs, and when I recognise them, I dislike them, because I don’t want my algorithm learning that this is what I want. It means building as rich a library as possible, downloading books in their original formats and buying physical books where I can, because we have been made aware of how, in the process of feeding these LLMs, hard copy books are being destroyed. Who will have the original knowledge sources in future? And it means being transparent. If I use AI for editing, as I have with this very piece, I can say so. If I use it in research, it is because I already have the research skills, and what I am trying to do is save time and have a research assistant I can afford, not outsource my thinking or my intelligence.
So ask yourself, where do you draw the line? What roadblocks are you setting up to guard yourself from losing your cognitive autonomy? Because yes, we are under threat. But while we are under that threat, we need to be honest. Honest about what these technologies are doing to us, and about the things we should never give over to them. But also honest about the world we created before they arrived. A world where there is increasingly no time to get lost, no time to think badly for a while, no time to walk around campus for three weeks until the idea finally comes. No time, sometimes, to just be human. If we want to talk seriously about our dependency on AI, we need to talk about both: the machine, and the world that made us feel like we needed it.
Anyway, those are my musings for this month. I’d love to know your own thoughts on this!
