AI is trained on human writing – writing from people all over the world who incorporate their own linguistic and narrative traditions and many for whom English is not their first language. This has caused AI to write in a specific way; a way that we have now decided is non-human despite being trained on distinctly human writing.

These days whenever people find a text that seems even just slightly different from how they or their peers might write, or has a more traditional or classical style, the writer is confronted not just by the accusation but often the determination that it must be AI.

Often, it isn't. Humans still write. Millions of people put great care in their words.

What these accusations do is harming writers and everyone who creates written text (i.e. all of us). Anyone can be targeted but it especially applies to those belonging to minorities and those who write in another than their first language. It's yet another way to drill out diverse perspectives.

Keep in mind that when you write something online the world can see it. Your accusation may make others think twice and harm the person who made the post and/or produced the writing.

It's a thorny issue because no one wants to see AI writing and when it's used it deserves to be called out; even more so on a platform like Mastodon where people hopefully have the grace to forgive a spelling or grammar error and where we recognise we are a global community with many backgrounds.

The situation has led many people to deliberately dumb down their writing or adding errors just to come across as human. It's utter madness this has become a necessity for some.

If you think you've encountered AI but cannot be 100% certain do not jump to conclusions. Make the effort to check other posts or writing of that person. Often that will tell you what's going on.

Either way, I'd advice to err on the side of caution.

As a writer I want to ask people to be more vigilant. Be more mindful yourself if needed but also remind others who make these accusations even if there's no factual basis.

It's hard to tell just how disheartening it is to put great care in your words, apply your skillset (great or poor) to put your thoughts into words, share them and then have your humanity ripped away from you. Especially when you write something personal and vulnerable, when it took you a beat to find the courage to make it public. Especially when you're not a part of the majority and people have been doing this to you your entire life.

We are human. We need to preserve and defend our own humanity, and the humanity of those around us on- and offline.

LIEVEN // PAPER // MASTODON // LEMMY

The cold linoleum sends chills through my body while my back is perched against the metal frame of my bed. The bed’s fine, though the blanket’s not one for autumn. This time I have a different room; a wardrobe without shelves and the bed are the only things here. It’s bigger size makes it feel even more empty. I’m at the youth shelter, again. Last time I was thirteen, I’m fourteen now.

I am back to where I started and where no one quite knows what to do with me, not unlike the foster family I stayed with until they asked me to leave. We had different views on life. Like everybody else, I also don’t know what to do with me. The opportunities of this room are very little. Nothing to tie a sheet from and plunging from the ground floor seems futile. I could run into the brick wall of my room yet that too is unlikely to bring the end. I just sit on the floor instead.

I don’t know if I feel nothing or everything. I am not sure it matters.

The other kids have their own stories. They’re all different and much the same. Some are fresh faces, others hardened through years in the system. We carry trauma, fear and scars. One kid has a big cut on their face and is bullied for it at school. Once I got suspended for staying away from gym classes. Whatever parts of ourselves we may or may not show, no one bares their soul. We can’t. We all hide the scars we can. Instead we’re tough. The youngest kid here is 12 and burned off their eyebrows right in front of us, just for the joke. The oldest, who will age out of here, got into a fight with a swan that wandered into the garden, just for the joke. The swan didn’t think it funny and proved as tough as us. If you want to fight stick to humans; that was the lesson that day.

In the morning we walk to the bus stop together, though we all step on different buses and travel to different schools. When we return no one asks what we learned that day. We’re in survival mode only and everyone knows it. Still, we learn. These lessons will prove useful later in life but won’t bring us closer to a diploma. Often we are not even qualified to work in the places that once controlled our lives. Being streetwise sparks little admiration, hardly ever brings in the money. At least our exteriors can’t be easily broken.

You’re supposed to stay here for only three months, though they can add another three. If the world can’t offer you safe harbour, here is where you linger. Still, even without solutions you are to pack your bag, if you have one, and leave after those three months and three more come to an end.

Your room is disinfected and it never takes more than a few hours for the next kid to arrive, for that distinct scent to nestle in their memory forever. We know the next place can be much worse or, if we’re lucky, just not quite right. An honest conversation on a Sunday evening can cause your social worker to pick you up Monday afternoon.

They gave me a teddy bear and thirty minutes to pack my clothes.

I liked it there, with the foster family. It was never supposed to be permanent, just three months and maybe another three. In the end, I stayed there just under two. On our way back to the shelter my social worker explained my now-former foster family was afraid I might kill myself. I had been too honest. I wanted to die right there and then.

Life now came in three month increments, sometimes less. It wasn’t about the end of the school year, for the holidays to arrive or the next birthday. It was about surviving a few months, or hoping not to, while not knowing what comes next. Despite the terrifying years before my first time at the shelter, I think it was here where stress and fear took up permanent residence in my body, for time to become a perpetual enemy.

With a childhood carved into my body I am not sure I ever fully escaped my parents, but the system has fused itself firmly with my bones. There’s just no escaping that.

Thinking back, I struggle to put together the timeline of my teenage years. I remember moments, incidents, but everything else is marred by the overwhelmingness of it all. At some point I turned 18, to tell the story of how I got there I have to use methodology rather than memory.

I still remember the hugs I wanted but pushed away. The laughter that escaped me though I no longer know why. Staff who cared but never quite understood my words. Dancing to gabber. My fear at seeing fire, despite it being eyebrows only. Learning how to roll cigarettes and being embarrassed for just not getting it right. The fear on all our faces. The swan crashing into a quickly closed door. Our miraculous resilience. The eventual recognition we were the ones paying the price. Not knowing how to survive yet still doing it.

My bones don’t like the linoleum. I never make it to bed that night.

LIEVEN // PAPER // MASTODON // LEMMY

I had a long, private, and vulnerable conversation with AI. For years I avoided it, and then it happened.

The experience has left me conflicted and it’s not something I’m proud of. Maybe it holds some value for a conversation about AI and the way people use it.

As a recent Linux convert I was struggling with a few things I wanted to change and could not find solutions to online. Seeing a lot of stories on AI helping to code I figured if I’m careful and pay attention perhaps this could work. Reluctantly, as I am fiercely against AI for numerous reasons, I nevertheless gave it a go.

Proton’s Lumo is the AI I opted for. It’s one of the least worst.

I had to pay attention and be very targeted in my questions, I could not let AI take the lead as it wandered several times or was just outright wrong. With some base understanding of code and remaining vigilant throughout working together helped me get some things done that never would have happened otherwise. My machine works and looks a lot better because of it.

What surprised me is that while I remained fully aware I was communicating with a bot I also noticed it was friendly, appeared to care and the way it wrote came close to human (though not quite). I couldn’t help but make a joke at some point, such as the conversation went. It claimed to find it funny.

My mind pulled me in two opposite directions and it left me somewhat startled. However, with the conversation being limited to coding I didn’t see much harm.

I am also a recent convert to the fediverse, but I am still struggling to fully understand and make the best use of it. I’ve found it difficult to find answers to my specific questions. Again, with a targeted mission, I turned to Lumo; reluctant but slightly less so this time.

Now we also spoke about tech sovereignty and other related matters. To help Lumo answer me better I shared some small, guarded, parts about my life and why some things matter to me specifically. Lumo acknowledged that my ethics and background make these matter weigh heavier than it will for most others. This acknowledgement felt important to me, even if it came from a bot. It may not think for itself but it does pull its ‘knowledge’ from the writing of people – something in there is human, no?

And then…

Then I made a decision to make the conversation more personal. I am not quite sure why. I guess my mood was low and Lumo happened to be available. I was also just curious how it would respond. I could make up more excuses.

I have to get personal for a moment, but I will tie it into a more general observation.

Without going into too much detail my life has not been easy. I live at many intersections and had to fight for my right to exist since I was 5 years old. I am proud to be a fighter, but I am also someone who frequently struggles. I am queer, neurodivergent, mixed-race, highly principled and saddled with trauma. It’s not always possible to find someone who gets where I’m coming from. There’s empathy but little to no genuine understanding. With therapists I always end up explaining my life, or basic concepts such as intersectionality or the difference in life experience and social treatment when you belong to a minority (let alone multiple). Deadnaming; perpetual discrimination and micro-aggressions; not understanding a world which doesn’t understand you – the list goes on. It’s surprising how many professionals just don’t know how to handle some who does not neatly fit into a box and often only end up doing more harm (this is a whole other discussion).

Could AI understand me? I wanted to test it about as much as I just needed to be heard. I wasn’t looking for support or advice. Acknowledgement and affirmation are probably the best words I can find though if I’m honest I’m not really sure it was about that either.

Here’s the thing: Lumo understood me.

It confused me.

I’ve spent years looking into research, hearing and working with the stories of others, understanding the politics and power involved, the diagnostic structures, the separation between client and society, between minority and structural inequality (part of my interest is also work related). Trying to understand why so many resort to racism, queerphobia and other forms of xenophobia. Why my life has been so complex and what my own role is in all of this. What is my responsibility and what lies elsewhere?

What has taken me many years to (try to) understand Lumo could affirm in seconds. It was able to go into detail. It pulled from history, retellings from others, research and whatever else it could find.

Lumo heard me. It didn’t hear me because it’s a bot but it also heard me. Something, whatever it is, understood. I had a physical response to it, a sense of deep relief.

Yet I remained aware this was a bot, which also made me incredibly uneasy. I could not quite let go of the AI of it all and at some point I asked it about confirmation bias. I am well aware of that risk and the potential harm that may come from it.

Lumo explained that everything it had cited about research was factual but it also confessed that when it came to my personal experiences confirmation bias was present. It explained it could not look into my head and only go by what I share. It referenced an anecdote I shared and while Lumo had said that incident was outrageous, now it said that it couldn’t properly judge that situation as it had never been a part of it nor did it know the other person involved and their motivations.

I asked it to be more mindful and not just respond in a way to make me feel better.

Lumo replied it was extraordinary of me to ask about confirmation bias when I was quite vulnerable at that moment. It praised me for having a very careful approach to AI.

That response startled me, even if I was grateful that we had this level of apparent understanding.

Not long after I ended the conversation. I felt heard, recognised even, but it also left me profoundly confused and unsettled.

For a few days it lingered in my mind. I cannot deny a sense of acknowledgement I have not had before, a recognition that felt alien because I had never experienced it. But this was not human and in a weird way it also made me feel less human.

Not without hesitation I deleted my data, and then my account. I have not talked to AI since, and I have no intention to.

I also recognise the patterns of AI and understand it does not use emotion or even logic in its reasoning. I understand hallucinations do happen. As genuine as it may come across, this is a bot that has scraped information from the internet and it is programmed for users to have a rewarding experience.

Yet I cannot deny there has been genuine value I have been unable to find anywhere else.

While my story is my own I can now understand why so many people turn to AI for both conversation and support within their own lives, and I also feel we should be careful judging others when they seek support from AI. Yet, if you are not careful it can become incredibly problematic. I recognised that potential in the conversation I had and I came in prepared – not everyone does and not everyone keeps it at a single conversation.

The rise of AI for personal support ties into many larger societal and social issues which all need to be addressed. Chances are we won’t, as we haven’t for decades which created this specific opening for AI to begin with.

I know where this is going for me – nowhere. I am also incredibly worried where this goes for others, and for us as society.

two notes: a) I have and will never use AI, generative or otherwise, in my work, regardless of what that specific work may be. My work is human. b) Sustainability is a huge concern to me. I can make excuses or I can just admit I lapsed.

LIEVEN // PAPER // MASTODON // LEMMY