News | Thursday, 28th March 2024

I know what you ChatGPT’d last summer

Why using AI to create content for your website could prove to be a big headache

David Edmundson-Bird, Principal Lecturer in Digital Marketing and Enterprise, and Faculty Lead in AI at Manchester Metropolitan University.
David Edmundson-Bird, Principal Lecturer in Digital Marketing and Enterprise, and Faculty Lead in AI at Manchester Metropolitan University.

David Edmundson-Bird is a Principal Lecturer in Digital Marketing and Enterprise, and Faculty Lead in AI at Manchester Metropolitan University. David is the Global Chair for judging standards at Don't Panic and a board member for futureeverything, an innovative global arts organisation based in Manchester. He also consults on digital communications for several small businesses.

Here, David shares his thoughts on artificial intelligence (AI), and using ChatGPT to create website content:

"Have you heard about Google changing their algorithm recently?

Google periodically updates the mysterious inner workings of its search engine, but the latest one is very significant and potentially going to impact a lot of small business owners who are embracing generative AI as a tool to help them market their products and services.

It all comes down to how some websites are using AI when creating content. This might feel like niche stuff but it’s already having an impact on businesses that rely on being found by customers in Google’s search results. 

Which may well include you.

The Algorithm and SEO

Google’s original search algorithm (named “Backrub”, which is not at all creepy) was developed in 1996.

Since Google’s launch in 1998, people have attempted to ensure their content (i.e. the words on their web pages) ranks highly when users type in search queries.

Over more than two decades, Google has tried to improve the quality of search results so that they are less prey to unethical and unscrupulous practices that seek to manipulate the algorithm. So Google and website owners have been locked in a digital arms race; there have been dozens of big annual updates and a constant series of daily tweaks to the search engine, and a sea of Search Engine Optimisation (SEO) techniques in response.

Much of Google's engineering effort has been to try and reduce the visibility of poor-quality web content that ‘games the system’. These changes have been designed to ensure that when you search the web, what you get is high-quality, valuable, and relevant content.

The trouble is, generating high-quality, valuable, and relevant content for your website can be time consuming, difficult, creative work that requires a lot of human intellectual capacity.

Enter Generative AI

Because content generation can be a pain, Generative AI tools, such as ChatGPT, Co-Pilot, Perplexity and Claude, with their ability to create compelling content quickly, and with almost no human intervention, seem like a great opportunity to make words on websites less of a chore.

If you’ve dabbled with one of these tools, it’s quite likely that you did so to generate exactly this kind of content; for your site, a blog, or a social media post.

The trouble is, unscrupulous operators have started using these tools to make vast quantities of content at very little cost or effort. And, they’ve been boasting about it.

So, why are bad actors making so much content?

‘Spammy’ sites draw visitors to web pages that contain lots of adverts, which provide a source of income for the site owners. They can also be used to lure people in for more malicious reasons, such as phishing for personal details or transmitting viruses.

Whatever the reason, manipulative players create vast amounts of shallow, lame content on many pages on networks of sites that they own – all targeting key phrases that people use to find useful content via Google.

Generative AI tools can produce vast quantities of shallow content on an industrial scale but with little effort and cost; this situation was inevitable.

Recently, users have complained that Google results to search queries have been… less than helpful.

Irrelevant material or content that doesn’t provide value is getting pushed to the top of results. This is a problem for Google’s business model, which wants to return trustworthy and relevant search results. People are questioning the value of the search engine's output.

What has Google said and done about it?

On 5 March 2024, Google announced a major algorithm update to get rid of a lot of this content by tackling ‘scaled content abuse’ – the manufacture of thousands of pieces of content to game the system.

The method they have adopted for tackling this is called a "Manual Action". This is a notice sent by Google to the owner of a website via their Google Search Console account that tells them they need to take action because Google has found a problem with a page on their website. 

In the case of pages that were identified as ‘scaled content abuse’, the consequences are that the pages in question have been made completely invisible (known as deindexing) from the search results – they no longer appear.

Researchers in SEO think that almost all of the pages affected by the 5 March statement used AI generated low-value content – and they estimate that half of those websites seemed to have content almost exclusively created by AI.

What does it mean for the business owner with a website?

It’s tempting to use AI to generate content for your website – for things like regular blog posts, information pages, and other content – especially when you are time-poor and busy running the business.

But using a tool like ChatGPT is not a panacea to produce valuable content. You risk:

So, what should you do with your web content?

Good things to read:

More news