Here's how I build civic tech, and how I think about AI.

I build software that helps people find coherence in firehoses of fragmented information.

This includes Wardroom, Radar, and campaign-infra used by several civic projects.

I develop and use machine learning models because I think they are genuinely useful for organizing information.

My concern regarding AI datacentres boils down to a lack of good governance. Regulations haven't caught up to what is essentially a new type of land-use. Questions around electricity, water, air pollution, and democratic oversight need answering.

FAQ

Do you use AI?

Yes.

Do you oppose AI?

No. AI is an ambiguous term for a large set of tools with highly disparate uses. Some of those uses are ruinous to society. Some of them I find genuinely terrifying. An equal number are unequivocally beneficial.

Why are you involved with/why did you start Stop the Datacentre(s)?

Upper levels of government have declined to regulate the physical infrastructure of AI. This means municipalities are the last thing standing between an appropriately wary public and a booming, unregulated industry.

Does AI decide what Stop the Datacentre(s) believes?

No. Stop the Datacentre(s) uses a combination of traditional software and (depending on the city) locally-trained and operated machine learning models to monitor public information sources and help volunteers keep track of developments.

These tools assist with information management. They do not determine strategy, write anything, or replace human judgment.