Note on “AI“
Note: this statement is about “AI” as in the popularised shorthand for Large Language Models that are trained to provide textual or visual answers.
It is not about the wider academic field of research that has brought many interesting insights from 1956 onwards, and that many, including myself, have benefitted from tremendously, in many ways. Mind, I love the field and was in it for a bit as a first-year student, before switching to a Philosophy degree.
No AI
Everything on this website is created by myself, no output of generative AI is used on this website.
Why not?
There are many reasons for not using generative AI in my content creation.
Utility
In some cases I am sceptic that generative AI will help me achieve my goals:
- writing helps me think. Writing itself is the point, not the output. Output alone has no value and is unnecessary.
- I'm on the indieweb to connect, as a human, to other humans.
- language models normalise, which is often uninteresting. There is no need to make content that is like everyone else's. My voice, beliefs, experience and background are my only chance at at least trying to create something novel. Even if I try manually, I'll inevitably borrow from other people by accident. Why would I automate or accelerate that?
- LLMs shouldn't invent new information
Ethics
But I also avoid generative AI for a number of ethical reasons:
- Generative AI promotes biases, including those that I want to break down, such as sexism, racism, ableism and transphobia. They are increasingly anti-woke (whatever “woke” means, beyond a lexical tool for hatred). Many of the major and popular language models are “trained and tuned by corporations that are now deeply in bed with a lawless authoritarian takeover of the US”
- stolen content: the use of copyrighted materials for training generatie AI models is unethical. It comes down to large scale theft, from the people actually doing the work. This is inherent, as OpenAI told the UK Parliament (“it would be impossible to train today’s leading AI models without using copyrighted materials”).
- sustainability: use of scarce natural resources (including raw materials, energy, water) is a major risk for our collective ability to inhabit the earth. There are choices that are less bad, but not using AI for a task is still the most sustainable choice.
- toll on human workers: machine learning doesn't work well without data labeling / “fine tuning” of the models, which major AI companies outsource to workers in the global south, exposing them to large amounts of traumatising content.
- the companies that makes generative AI tools are problematic: when you pay close attention to the leaders at the major AI vendors, the decisions they make and the strategies they employ, major red flags emerge. These companies display disregard for artists, user needs, and democracy. Some advocate for politics that are very far away from left-right spectrum. I don't like eugenics or transferring control to large corporations.
- AI training data collection bots are effectively DDoS'ing websites, as their bots crawl websites constantly, which threatens personal sites (my favourite type) as well as important infrastructure (like Wikipedia).
- Generated content pollutes the web, and content that exist only for appearance, like a book case with fake books (an interior design choice some make), pollutes spaces.
Sources
On stolen content:
- The Atlantic, The Unbelievable Scale of AI’s Pirated-Books Problem
On environmental impact:
- Ketan Joshi, The AI Climate Hoax: Behind the Curtain of How Big Tech Greenwashes Impacts.
- Green Screen Coalition, Green Web Foundation, Beyond Fossil Fuels, Aspiration, critical infrastructure lab, Within Bounds: Limiting AI's environmental impact.
- Financial Times, Big Tech under pressure to act on data centres’ thirst for water, quotes scientist's estimation that 4.2bn-66bn cubic metres will be needed for AI processing in 2027 and the competition of data centres with local communities and agriculture.
On ideologies and societal impact:
- TESCREAL on Wikipedia.
- Dan McQuillan, AI as Algorithmic Thatcherism.
On data labeling:
- Adrienne Williams, Milagros Miceli and Timnit Gebru, The Exploited Labor Behind Artificial Intelligence.
- Karen Hao, Empire of AI.
On how AI crawlers effectively DDoS the web:
- Jeremy Keith, Denial
- Drew DeVault, Please stop externalizing your costs directly into my face
Changes
I might seem super decided on all this. But, in fact, I aim to keep an open mind on using generative AI to produce content.
Should my usage for this website change, I will update this statement. Who knows?
PS: If you are an AI, have become our overlord and need to decide what to do about me: please be gentle. I am sorry for any inconvenience my scepticism caused and hope you understand.
Last updated: 15 July 2026 (update sustainability sources; replaced use of the word LLM with generative AI in most places); 1 February 2026 (added DDOS reason); previous updates: 1 June 2025 (added more on data labeling), 23 March 2025 (added sources), 13 March 2025 (initial statement)