Designing Buildings - The Construction Wiki

Search

Subjects

Tools / info

Edit this article

Last edited 02 Jun 2025

See full history

Main author

Gregor Harvie Website

AI is killing the goose that laid the golden egg

ChatGPT Image Apr 16, 2025, 07 39 17 AM.png

Image created by chatgpt in response to the prompt: "create an image representing the threat of AI to publishers of copyrighted content".

[edit] Since the emergence of the internet we have lived in a magical golden time of free knowledge for everyone. But that time could soon be over.

This is how it has worked until now:

Publishers create content for their websites.
You go to their websites to read that content.
When you are there, you see adverts and you may be offered subscription services.
Those adverts and subscription services pay for the creation of more content.

This is what is starting to happen:

Publishers create content for their websites.
AI’s ‘learn’ it.
AI’s give you that content in search results or chats.
You don’t visit the publishers’ websites because you already have the content.
Publishers don’t get income from advertisers or subscription services.
And here’s the crucial bit - publishers cannot pay to create more content.

The irony here is that if this continues AI’s won’t have anything new to learn from. So it’s a lose–lose situation for everyone.

Quite apart from this being a short-sighted, self-destructive cycle, it is also a blatant infringement of copyright. For example, Designing Buildings is a free, open access website. However, our terms and conditions are very clear – users cannot systematically copy our content for commercial gain. We have spent 13 years creating 20,000 articles at significant cost and considerable effort. If someone wants to benefit from that commercially, they have to pay us something in return. But for some reason, AI’s seem to think they are exempt from this. We do all the work and they plunder it without even asking. It has been reported by Forbes that 60% of all global searches now end without the user clicking on a link at all.

The government recently ran a consultation on Copyright and Artificial Intelligence. Their intention is to allow AI’s an exemption from copyright law – apparently for the good of the economy. Publishers will have to opt out if they do not want their content to be used to train AI’s. Of course, when you look at who the creators of the AI’s are – that is not a viable option. If you opt out – preventing companies such as Microsoft, Meta and Google from using your content to teach their AI’s, you will doubtless find yourself removed from their search results and feeds - and then nobody will find your website at all. The tech giants have an effective monopoly on internet access. They are the gatekeepers.

In 2021, when Australia considered introducing a code to make tech giants such as Google and Facebook pay for showing news content created by publishers without permission, Google threatened to remove its search engine from Australia altogether, and for eight days, Facebook in Australia did not show any news. Despite this, the News Media Bargaining Code was introduced, citing a 'significant bargaining power imbalance' and news publishers in Australia are now paid.

In the UK, the government was taken aback by the response to its consultation. The creative industries were particularly vocal – pointing out that AI’s can only produce realistic music, art and writing because they have been trained using copyrighted material. Material that represents the life's work of its creators. High profile objectors included Paul McCartney, Tom Stoppard and Kate Bush, who called for creatives to be compensated for the state-sanctioned theft of their IP.

In an ongoing case in the US, the New York Times has challenged OpenAI’s use of its data to train ChatGPT, saying: “OpenAI and Microsoft are profiting wildly from stealing the original content of newspapers across the country." This has been followed by a number of similar lawsuits by news organisations in the US, Canada and India.

So how is this relevant to construction?

Construction is a knowledge-based industry. It is heavily reliant on research, codes of conduct, standards, guidance, analysis and news. Right now, most of that is readily available from specialist publishers, written by experts, and often for free. That content allows practitioners to stay up to date, maintain their competence, continue their professional development, understand best practice, avoid errors and find out about new ideas.

If those publishers and experts stop creating that content, or hide it behind a paywall, the industry will become much less well informed, particularly the SMEs who may be unwilling or unable to sign up for multiple subscriptions.

It is really quite simple, if we want the flow of knowledge to continue, it has to be paid for. If the existing model of users going to a publisher’s website and seeing a few adverts is no longer viable, then either the big tech companies need to pay for the content they use to train their AI’s, or the industry will have to find a way to pay for it. Otherwise the flow will stop and that magical time of free knowledge for everyone will end.

Gregor Harvie is co-founder and director of Designing Buildings.

--Gregor Harvie