How to save your writing from the AI robot apocalypse.

This post was written a while back now, deep time in AI terms. It was an exploration of some AI apps, trying out writing copy for the first time on them. Since then, I've learnt a lot more about the subject, so it's time for a rewrite.

I see I was naive, re-reading the previous version of this post. Here are some of the things I know now which I didn't then.

In this post :

• Tech companies quietly use your writing to train their LLM models
• The impact of AI on Education
• The impact of AI on web search
• I am surprised how much I dislike AI generated content
• The impact of AI on writers
• How my thinking on LLMs have changed

Tech companies quietly use your writing to train their LLM models

I didn't know that many companies would use peoples' writing to train LLM without being very upfront about permissions. I write here about how to protect your online writing.

Companies like Meta make it difficult to refuse this training. Substack, LinkedIn, ChatGPT, Google, they have all sucked up your writing. Switching off that permission now is not back-datable, apparently.

In one of the most expensive legal arguments written on the back of a packet of Benson & Hedges cigarettes, Meta have stated "it's alright because everyone else does it."

The impact of AI on Education

Both in a good way and a bad way. Students have embraced using LLMs to write essays. The essay is an important tool for thinking - writing is more than just putting words on a page. If you outsource this it impacts learning.

On the other hand, companies like Khan Academy (and their AI tutor Kanmigo) show how AI can be used to facilitate learning.

But in the sector, nobody is quite sure how to square the circle yet.

The impact of AI on web search

The environmental cost is considerable.

I couldn't forsee at the time how search would change - that AI search would take text you had written and repurpose it, then deliver that as the result.

I had a strange AI search experience which I wrote about - what it feels like when the SEO works.

I am surprised how much I dislike AI generated content

I don't think it's particularly good at writing, it's kind of flat and slightly uncanny valley-ish. It's not good at subtext and... it doesn't feel. It can pretend to be a poet, but it doesn't feel.

I understand how AI generated video is popular but I find I have a bad reaction to it, so much so that I have had to stop looking at Instagram.

The impact of AI on writers

I had an inkling that this tech would have a very strong effect on copywriters and content writers. That was hardly a brilliant insight.

What I didn't realise was how immediate it would be. I write about the devaluation of labour in the post 'Where will young writers learn their craft.'

I also didn't know that companies like Meta - Meta! - would use pirated material to train its LLM on. You can search here to see if your book is included in the database they used, This is on the Atlantic website - Search LibGen, the Pirated-Books Database That Meta Used to Train AI. One of mine is.

Here is another quote from an Atlantic article called The Company Quietly Funneling Paywalled Articles to AI Developers.

The common crawl foundation is little known outside of Silicon Valley. For more than a decade, the nonprofit has been scraping billions of webpages to build a massive archive of the internet. This database—large enough to be measured in petabytes—is made freely availablefor research.

In recent years, however, this archive has been put to a controversial purpose: AI companies including OpenAI, Google, Anthropic, Nvidia, Meta, and Amazon have used it to train large language models. In the process, my reporting has found, Common Crawl has opened a back door for AI companies to train their models with paywalled articles from major news websites. And the foundation appears to be lying to publishers about this—as well as masking the actual contents of its archives.

I am also worried that there will be less opportunities for young writers to learn their craft. In the same article I reference Writers' Guild and ALCS surveys of writers and how much they know about how their work is being used.

The ALCS Chair, Tom Chatfield, wrote about an ALCS’s survey of over 13,000 writers (N.B this link opens a pdf), 

77% don’t even know if their works have been used to train AI systems. Among those who do know, only 7% gave permission. And a remarkable 91% felt they should be asked for permission to use their works.

How my thinking on LLMs has changed

I'm still positive about these tools and I use them to help build a frame of reference when learning. I also use them when I'm doing something on the tech side. I get them to re-write a bibliography according to Harvard and stuff like that.

But the time of blindly trusting the technology has gone.

Here is a short video with former Google CEO Eric Schmidt on the future of AI.

LLMs are just a small part of the picture. This documentary on Google Deep Mind called the Thinking Game is fascinating, thought-provoking, and illustrates it's only one direction we're going in.

I write about SEO and how to help people with their online writing, but I do it from the writer's perspective and try and highlight what writers should be careful of. If you are interested, the pillar post is SEO for writers, creative and arts organisations.