What is Google’s stance on AI-generated content?

With the rise of AI tools, we’re increasingly asked about what Google thinks about AI-generated content. Can Google recognise it? Will we be penalised - or even rewarded - for publishing it? What will happen to my rankings if I use AI tools? These are all great questions.

Here, we explore Google's perspective on AI-generated content and discuss what this could mean for your SEO strategy.

Google’s position on AI-generated or AI-assisted content has been laid out very clearly: appropriate use is not against the rules. Google updated its guidelines on helpful content last year, changing “content written by people” to “content created for people”, which many believed was a nod to this stance.

In the documentation, Google states it is not against AI-generated content and not all ‘automation’ is classed as spam. The ‘appropriate use of AI or automation’ is the key element - the spam policies are clear, and the content has to be helpful. It doesn’t mean that Google favours this type of content either - “using AI doesn't give content any special gains” - but there’s no penalty for using it if the content is high quality.

Over the years, Google has constantly responded to people using different tactics and technologies to “game” search results. Content has always had a big part to play, with tactics such as ‘article spinning’ being one example where large amounts of lower quality content is produced to influence search engines.

‘Article spinning’ involves rewriting the same piece of content multiple times with slight adjustments, meaning that huge amounts of content can be generated quickly. Often, these ‘unique’ articles are provided as guest posts for other websites, or published on private blog networks to generate links. It’s an old, spammy tactic that results in articles that aren’t helpful to read and provide no value. Huge numbers of articles were generated in this way and the Google Panda update in 2011 was a major move in tackling this type of poor content.

Google’s systems have been continually adapting to counter tactics like this, and generally we see cycles of new spam techniques, and then Google responding to them over time; typically with algorithm updates. AI-generated content is not the first time that content has been automated and therefore Google has stated its systems and spam policies are already set up for this. If Google has been tackling poor content for a long time, poor quality AI-generated content is just a similar challenge in a new guise.

It doesn’t mean that Google always gets it right, though. The 2023 AI-powered equivalent of article spinning is scaled content abuse. There has been a huge amount of coverage and criticism from the SEO industry, and the wider media, about the amount of AI-generated, spammy content in Google search results over the past year.

Scaled content abuse is the mass creation of pages with the purpose of manipulating rankings in search, rather than helping people. This can be done using a variety of techniques, but the most relevant is using Generative AI. Using AI tools to create pages for SEO purposes, rather than to add value for real searchers, is against the guidelines. Google’s response to these practices was the March 2024 core update, which was a complex systems update and the introduction of new spam policies. The aim was clear: reduce the amount of low quality content in search results.

The rollout of this core update completed in April, and many sites have seen fluctuations in their visibility in relation to this - it was a complex update that affected multiple core ranking systems. As of the 26th April 2024, Google states that 45% less low-quality, unoriginal content can be seen in search results due to the core update in March. They are taking targeted action on sites that go against the spam policies, resulting in an increase in recent manual actions.

Do not use AI tools to mass-generate content for your website, especially if this content is exclusively designed to improve rankings in search engines - it is not a good idea and it won’t work out for your organisation longer term.

The core update in March 2024 also marked a major milestone in how Google systems identify content helpfulness. To assess the risk of creating AI-assisted content, it’s useful to understand Google’s overall position on generating helpful content. Helpful content is a critical concept for success in Google search. Since 2022, the concept of helpful content has been a major overarching message from Google about how to rank well in Google search.

The focus is on people-first content: don’t do anything for the search engine, do it for the people behind them.

The messaging from Google on quality, helpful content has been highly consistent. If you want to responsibly use AI to assist in the creation of high-quality content, that’s not against the guidelines. Ultimately, you need to focus on creating content that meets the needs of the real people searching, and ensuring the content is accurate, reliable, and trustworthy.

We would recommend leaning on the experts within your organisation to ensure that content is appropriately reviewed, consistently updated (where appropriate), and include clear authorship information. When working with nonprofits, we almost always find an enormous amount of unique, internal expertise that should be leaned on when creating web content. Expert authorship or review processes should be included on pages where applicable.

Where it would be useful for a person reading, you may wish to include an AI-assisted authorship statement. Google includes this in its helpful content guidelines - where you’ve used automation to substantially generate content, is the use of automation evident to your visitors? Are you providing information on how automation was used or why it was useful? These are things you might consider including to demonstrate your overall reliability and trustworthiness.

At Torchbox, our approach is that SEO shouldn’t be separate from your overall people-first content design, it should be a core part of it. We do very little ‘for search engines’ and everything we can for the real human beings who use search engines every day. If you follow the same responsible approach when adopting new technologies and creating unique content, you’ll be in a strong position to maintain and improve your search performance.