Reading Time: 2 minutes

I have added a sentence to the end of my Privacy pages which states the legalese was written with the help of an LLM. Not because I worry people might be duped by the content, which is standard and perfunctory, but because I think, collectively, we need to be more honest about LLM usage. At the moment, it can feel like admitting to using an LLM is akin to being caught doing something unspeakable in public – which is silly. And people are using it surreptitiously, which is not helpful. I have used a Thesaurus for as long as I can remember, and most readers will be extremely grateful I also use a spellchecker. Why not an LLM to copy edit too? (I am sure there will be plenty of answers from those who despise all things LLM-related.)

This is not to understate the significant shifts humanity is facing, nor the various dangers and threats that dynamic pattern recognition processes pose (or rather, their billionaire owners’ lack of care pose). I am glad to see EU rules coming into force tomorrow, which state that companies should indicate when they have used an LLM or image generator to produce content; and as good practice, I have done so too, even though – being in the UK – I am not subject to EU legislation, and the directive is aimed at more worrying fake-news content than Privacy notices.

Screenshot-2026-08-01-at-18.03.31.png

Some will say it’s not enough or that the legislation is self-defeating. Perhaps. But I doubt we got it right with motoring legislation to begin with, as an example. It needs to happen, and it is good the EU is implementing laws and promoting good practice, whereas the US government and the billionaire owners are resisting constraints designed to protect society for all they are worth, which is an awful lot, as we are all too aware.

The Guardian also had something to say about the new laws in yesterday’s paper.

And here is an image made by a dynamic machine processing in 2023 based on a small collection from a book about bodies and gynaecology that informed an earlier work (Lamella:// (2023) which drew on Zabriskie’s Obstetrics for Nurses (1960)).

Added 12/08/26: Individuals from the industry begging for legislation to be put in place
https://www.theguardian.com/commentisfree/2026/aug/11/openai-anthropic-google-deepmind-letter
https://www.pacingthefrontier.com

Discover more from Field Notes

Subscribe now to keep reading and get access to the full archive.

Continue reading