Communities are confident on where they don’t want data centres, but it’s a little more complicated on what the data centres are for

My lunchtime shenanigans were a little different yesterday: I tagged along with EG Nefarious to a rally against the use of AI in Australian Education. I was very happy to be there, and probably one of the loudest voices. It’s not our first protest or rally; I’ve been ‘supervising parent’ with the kids from the first School Strikes for Climate to the more recent Herzog Protest (AKA NSW Corrupt Bully Police Force in Action). We research beforehand, read the support material, and have plenty of intelligent debate at home before we step outside. I am a big believer in community action and lifting our voices to be heard. I’m there to support the kids, and add an extra pair of watchful eyes for if things turn nasty.

AI in our schools and universities is both a community issue, and a deeply personal one. EG Nefarious finished his senior high school exams last week, with plans to attend university. It was hard enough in high school with the push of AI, but the volatility of the topic at university does not ease his concerns. This is a kid who feels strongly about science, education, and critical thought. He sees the complexity of the issues surrounding AI, and he wants to trust the discussion. But at the same time, there are too many conflicting interests, too much money and greed, and the only way he sees any chance of his concerns being heard is to attend rallies and moratoriums, like today.

Artificial Intelligence: We Actually Know Very Little

First up: No generative AI was used in the creation of this article, or any article that I create. It’s all my own work, tapping away at the keyboard, constantly deleting my mixed up words, and desperately trying to pull it together into a coherent argument to encourage you to think for a few more minutes.

I used a key term in the paragraph above: Generative AI. Because one of the things I learnt from the rally today (and have thought about a few times earlier) is how little we actually understand about Artificial Intelligence as a whole society. It does not help when the big tech CEOs and our leading governments keep changing the terminology, despite years of collective understanding and research amongst the people who actually created it.

Most of the AI we hear about today can be split into a few categories based on how we use it. For example, generative AI can generate content in response to a specific prompt or request. Sometimes, AI is categorised by its capabilities. It can have a narrow list of expectations and parameters, like face recognition and algorithm recommendations. Or it can use a greater learning capacity, using memory recall and prediction to be used in things like medical analysis. To put it bluntly, not all AI is bad. But when you use targeted terminology to change the way we think about it, then we start having issues.

For example, ‘Generative AI’ is often described as ‘generating original content’, as per IBM’s own definition. This is bullshit, because AI is yet to learn this skill without first relying on learning models. Write me an essay? Sure, let me scrape these websites for information and language style. Create a political cartoon for me? Okay, let me first get an idea from these other websites. Analyse this medical scan and tell me what urgent care is required? No problem, based on a million other studies and reports I have been trained with. There is no original content with Generative AI.

But again: not all of it is bad. If that medical scan is going to save a life, I’m happy to support doctors in whatever way works. Just as an example. The thing is, big tech CEOs have figured out the best way to make money out of this, and it’s not with those medical reports.

AI in our Education

During the rally, one of the speakers mentioned the use of Turnitin and its relationship with students. Turnitin (pronounced ‘Turn It In’) was initially promoted as a peer-review online system and has evolved into an internet-based anti-plagiarism service. It has always relied on the submission of papers from students and academics to supplement its database and provide the basis for its comparisons. In 2023, it included a feature it claimed could detect AI generated content, like ChatGPT. Universities and schools loved this, with the feature used across 16,000 institutions worldwide with 71 million students enrolled with the service. That’s where the money is: recurring subscriptions that lock customers into using the AI that claims to solve the problem of students using AI. However, many universities are now cancelling their subscriptions due to questionable policies with Turnitin (read more over at the Queen’s Journal here).

The problem with Turnitin is two-fold. First, it’s heavily unreliable. Like, serious recurring complaints. If a student used Grammarly to assist with grammatical corrections for an essay they wrote themselves, it would still flag as AI. Counterproductive in situations where teaching staff are already expressing bias just by doubting the student’s work. And don’t even get me started on the racial profiling for students whose first language is not English.

The second issue is copyright. This year, Turnitin proposed to change its licence agreement to allow Turnitin to use student work to train its AI models. As you can imagine, students and academics were not impressed. After reading the room (nice change), Turnitin has postponed any changes to its policy until 2027. In the meantime, they have opted to explain their AI categories; specifically the difference between AI Classifiers (how it assigns a numerical score for the likelihood of AI-generated papers) and Generative AI, the latter using the same misleading definition as the IBM example I mentioned earlier.

Put the two issues together and you are faced with a new problem: students’ papers used to train the Turnitin AI models are going to set a particular standard of language style, research, and presentation. When we are looking at hundreds of students with the same assignment topics or questions, you are likely to have overlaps. That increases your chance of false AI claims, which in turn causes more problems with your students.

Nevermind the fact that Universities HATE anything that interferes with their potential to claim copyright on research. Oh sure, it’s about the students’ IP but Universities will not share that possibility with anyone, and especially AI companies like Turnitin.

Are AI Data Centres the Problem?

I am not a fan of data centres, especially in residential areas or within close proximity of schools. Let’s be serious: there is no reason for either of those. There are, in fact, a few reasons why we need to reconsider both the quantity and the location of so many data centres.

The rally today highlighted the recent community win against building a data centre at Cooks River in Arncliffe (Sydney). The developers had originally proposed building a “logistics precinct” on the rezoned Kogarah Golf Club site, including an AI data centre. It was conveniently located right on the Cooks River, sparking an environmental concern. After protests and complaints to local authorities, the developers will continue with the logistics hub but have cancelled the AI data centre. It’s one of many community wins this year, following Lane Cove (NSW) and Fitzroy (Vic).

Data Centres have now become the poisoned chalice of policy, and no politician wants to be dragged into the debate if they can avoid it. Even Nationals leader, Senator Matt Canavan is afraid to discuss it. On ABC’s Insiders (8 Nov), Senator Canavan was happy to voice his support for a nuclear powerplant in his Capricornia neighbourhood, but when asked about a data centre, he was suddenly very eager to point out the need for community consideration and assessments for better locations.

The hard truth is that we do need data centres, but we shouldn’t be supporting the data centres designed purely for training Generative AI, including Large Language Models. Good luck with transparency on whether that’s happening. We genuinely need data centres for places like medical facilities and banking institutions, even government organisations (though I have my own humanitarian law views about that). The further the data centre from these organisations, the greater the risk of corruption or interference or simply delay in relaying the information needed. This is called latency and latency is a big issue.

Even if the data centre is further away, it can still cause issues for the common people. Places like Fitzroy have shown that it’s not always the data centre location that is the problem. For those residents, they were facing oversized electricity poles and cables to manage the increased power needed for the data centre outside their local neighbourhood.

If Data Centres aren’t the Problem, What is?

And now we’re back at the Big Tech CEOs and government leaders.

Why are they pushing so hard for everyday people to use AI all the time? Why is Google coercing schools to increase their use of AI tools, or risk losing Google Search completely? And why are schools including AI prompts in their assessments for students?

That last question is a serious issue. EG Zaltu was recently instructed by their Visual Arts teacher to “use AI (like ChatGPT) to generate an essay response to the assessment question. Then write your thoughts about the AI report.” Let’s just sidestep how deeply disappointed I am in a Visual Arts teacher supporting the use of AI in the current social environment of copyright and creative theft. There are many issues with the use of AI in schools, including the effect of cognitive offloading and critical thinking. Yeah, because that isn’t already an issue in society. But if we can hook the kids on AI at school, they’ll be paying customers as adults!

The push for AI adoption, and so-called ‘need’ for more data centres is about economics. Many businesses simply want more productivity with less costs, primarily people. Very few are asking the harder questions: where is this data coming from, where is it being stored, who is fact checking, what inherent bias is included, etc. Governments are also benefitting from AI, both in the workplace and in the community. There is a lot of money involved, and in the current economic environment, everyone is looking to save a buck. Often without realising they paid it in the first place.

That’s where community action is important. It forces the discussion, slows the movement, and gives us a minute to ask what is going on behind the money. Success stories like the Cooks River data centre slow down the building-rush and force authorities to look at the greater impact: environment, residential development, health, and genuine enjoyment of living. Stories like Turnitin force institutions to slow down and think about their students (shock!), as well as intellectual property and the amount of money spent on subscriptions. And rallies, like the one today, force us to individually slow down and think of our AI consumption. Right now, AI looks fascinating and fast and furious. But it doesn’t have to be. No matter how much the companies want us to believe otherwise.

If there was one thing I took away from the rally today, it would be the importance of seeing the Big Picture with education. That’s the value of learning. I know, that sounds so cringe but it is pretty much the biggest point. One of the speakers mentioned a graduation ceremony at UTS earlier this week, which included one of his PhD candidates (congratulations, Doctor!!). During the Chancellor’s speech, the Chancellor raved about the use of AI, and talked up all the benefits. Including the fact she had used ChatGPT to write her speech. She wasn’t the first; we suspect EG Nefarious’ school principal did the same for the Year 12 Presentation Night. In both cases, most students felt disrespected by their so-called authority figures. Neither the UTS Chancellor nor the high school principal could be bothered to make the effort and write a speech that celebrated their students. By using Generative AI, these academic professionals made their speeches about themselves. The Chancellor and the Principal clearly learnt nothing about their students, and showed the students they didn’t need to learn anything either.

Slow down the data centres. Slow down your AI use. And keep AI out of our schools. Give our kids the chance to develop their critical thought. Give them time to learn how to express their creativity in all disciplines. Most importantly, give our teachers more time and resources to connect with kids, not by outsourcing to AI but with government support.

Maybe… just maybe… we can get our real creative generation back.

—

If you are interested in joining the community action against AI data centres, the next moratorium is 11am Saturday 5 December 2026 at Parramatta Square (NSW). There are meet-ups happening all around Australia, so if you’re not Sydney-based, find your closest. Even if you don’t agree with everything, if you at least agree with a slow-down then make your voice heard.

Make Your Voice Heard

Interesting topic? Check out my other articles about AI under the menu heading “Think: If I Ran the People Zoo“

Categories: If I Ran The People Zoo Think

Tagged as: AI community action data centres education environment protest schools science university

Evil Genius Mum

Evil Genius Mum
- Taking over the world, one blog post at a time