Mastodawn

Here's some of my thoughts on the new Radboud "Strategy & Ambitions". A university's strategy MUST be determined by, and widely supported by, students, professors and other employees. Our key value used to be "sustainability": You have a part to play! Sure, advertisement bureaus were paid, cringe movies were filmed, social media campaigns were launched, but at least sustainability was something we all can support. https://www.ru.nl/en/you-have-a-part-to-play
🧵 1/7

You have a part to play. | Radboud University

With You have a part to play, Radboud University drew attention to sustainability. Read about this campaign and contribute to a more sustainable world.

Show thread

Djoerd Hiemstra 🍉1d ago

That changed! Our key value sustainability is now "sustainability and AI". The strategy repeats the narrative from Big Tech. AI is "developing rapidly and penetrating more and more areas of society and science", "AI directly impacts our region and society", and AI "can help us make our processes more efficient and enhance the quality of our services". https://www.ru.nl/en/about-us/mission-and-strategy/strategy-and-ambitions/sustainability-and-ai
🧵 2/7

Sustainability and AI are and will remain important topics | Radboud University

AI and sustainability are and will remain important topics for us as a university.

Show thread

Djoerd Hiemstra 🍉1d ago

This narrative, "AI changes everything", is problematic. It's the narrative put forward by OpenAI, Anthropic, Microsoft, Facebook, Amazon, Google, and let's be clear why: If you invest trillions in new technology without any clear return, of course, you would have to claim that it changes everything. In reality however, more than a trillion Euro is invested in autonomous agentic chatbots that are correct slightly more often, and that do no longer persuade people to commit suicide.
🧵 3/7

Show thread

Djoerd Hiemstra 🍉1d ago

Let me stress, AI as a discipline is important. I consider myself an AI researcher. I wrote my PhD thesis about language models. Some of my colleagues pragmatically embrace the new strategy (more money for us!). Not me, I'm ashamed. Is AI research more important than cancer research? more important that particle physics? more important than language studies? A university board should never even be tempted to make those decisions. Strategic decisions should be made by us, from the ground up.
4/7

Show thread

Djoerd Hiemstra 🍉1d ago

Some of our most talented researchers in AI, linguistics and computer science (they're on the Fediverse if you need them) oppose the AI hype. The board and the consultants who wrote the strategy may have listened to these researchers and claim the strategy is a compromise, that it promotes only "value-driven" AI, that we "consciously choose", etc. However, any strategy that is a compromise is a failed strategy. A common strategy must have consensus. Is there consensus? Clearly not.
🧵 5/7

Show thread

Djoerd Hiemstra 🍉1d ago

And then there's the elephant in the room: AI of the chatbot and generative kind needs large amounts of energy. Letting a chatbot check the grammar and spelling of your text costs a million times more energy than the old Word grammar check (I am generous here, likely hundreds of million times). It's a completely UNSUSTAINABLE technology. Data centers are being built everywhere that strain the electricity grid and use water for cooling. AI is a big threat to the energy transition.
🧵 6/7

Show thread

Djoerd Hiemstra 🍉1d ago

To conclude, I feel ashamed by the "AI and sustainability" strategy of the university, I cannot support it. The consultants that wrote the strategy missed two red flags. 🚩1 Some of our most talented professors object! A university strategy should be based on consensus. If your strategy is a comprise, forget it! 🚩2 AI is a horrendously unsustainable technology taking all the surplus electricity we have. You simply cannot have both AI and sustainability: Make up your mind, Radboud University!
7/7

Show thread

Ed.ucator 1d ago

@djoerd I coudln't agree more, thank you.
Plus: if a university wants to profile itself in 2026 and chooses AI to be a discerning characteristic, their view of the research field is rather clouded

Show thread

Djoerd Hiemstra 🍉1d ago

@EdwinvMeerkerk thanks, yes, clouded indeed, and in the cloud as well.

Show thread