Google’s guidelines for using AI-generated content on your website

Plus, Kontrolit’s policy on using AI in our services, and our best-practice recommendations to you.

Since the huge sign-up experienced by Chat GPT in the first few months of 2023, everyone’s talking about AI. That mass appeal is now prompting a more serious discussion: how AI tools should be used responsibly, in ways that will help rather than harm your website.

We’ve been asked a number of questions about AI tools recently, and in particular Chat GPT:

• Will Google penalise AI-generated content?
• What are the do’s and don’ts of using tools like Chat GPT for generating website content?
• What is Kontrolit’s policy for using AI tools?

Will Google penalise AI-generated content?

Google has been clarifying their message and guidelines around AI-generated content for some time, but the simple answer is: No, it won’t automatically penalise you, but there are strong caveats and guidelines to be aware of.

1. Google’s main principle remains the same as ever: content must be valuable. Demonstrate Expertise, Experience, Authoritativeness and Trustworthiness (these are Google’s core EEAT principles).
2. Its ranking algorithm will continue to reward quality content and weed out spam.
3. Google says: “Our focus [is] on the quality of content, rather than how content is produced”.

Here is Google’s official guidance about AI-generated content from their February 2023 blog post.

What are the do’s and don’ts of using tools like Chat GPT for generating website content in a safe and responsible way?

Google is not giving everyone a green light to churn out copy-and-pasted Chat GPT-created content for their website.

AI tools can streamline your workflow and improve your output, but they have limitations; they aren’t always accurate or up-to-date, even though the text can sound convincing. Chat GPT, for example, sometimes generates wrong answers, has limited knowledge of world events after 2021, and has been known to produce biased or harmful content. 

The do’s

1. Make sure you’re giving your readers something new, with value. However good your AI-generated content seems on the surface, does it conform to Google’s EEAT principles (as mentioned above)? Does it compare favourably with a top-ranking competitor’s content?
2. Use Chat GPT as a starting point to help generate ideas for your content, and then build on those ideas with your own knowledge and research. For example, ask it to suggest a structure for an article, with suggested headings. Even if you’ve done some research yourself, there may be ideas you haven’t thought of covering.
3. Proofread and edit your content carefully to make sure it's accurate and makes sense. AI models do not always generate perfectly polished content, and expertise may be lacking.
4. Revise the content it generates to make it your own. This will help you avoid duplicate content and make it more unique. It also ensures you achieve the tone and voice that aligns with your brand and target audience.
5. Experiment with different prompts in Chat GPT. With practice, a well-structured prompt can generate a response that’s much more useful.

The don’ts

1. Most importantly, don't sacrifice quality for quantity: While Chat GPT can generate a lot of content quickly, it's important to always prioritise quality content over “filler” content that lacks value.
2. Don’t take Chat GPT’s output at face value without checking its accuracy. It never says, “I think…” or “it might be…”, which can give rise to a misguided confidence. 
3. Don’t rely on Chat GPT for all your content needs: it can be a helpful tool, but your audience can tell if the content is not unique or lacks a personal touch.
4. Don’t use the content without reviewing it first, to make sure it's well-written and coherent.
5. Don’t assume that AI-generated content is optimised for search engines. Tools such as Chat GPT aren’t up-to-date with keyword trends or search volumes, so you may need to rewrite content in line with your own keyword research.
6. Don’t use any auto-generated content without first checking that it hasn’t been plagiarised. The problem with Chat GPT at the moment is that it doesn’t quote its sources, so you can’t be absolutely sure that the content it’s suggesting hasn’t been scraped from another website.

Google hasn’t clearly set out what a penalty might look like, but as with any shoddy content it comes across, the very least you can expect is for Google to ignore it. It certainly won’t rank well, and there may even be a manual penalty if Google thinks you’re using AI purely for manipulating rankings.

There is likely to be a blitz of lazy, low-quality content published by marketers who are using AI tools in the wrong way. Skilled researchers and content creators who can rise above those methods will get the most benefit from platforms like Chat GPT.

What is Kontrolit’s policy for using AI tools?

In our world, new technology, software tools and Google guidelines are launched with head-spinning regularity, so Kontrolit is very capable at adapting our methods to benefit our customers.

We see opportunities to use AI in various ways for SEO, PPC and web development, but our overriding policy has always been – and will always be – a completely ethical approach.

• We do not use AI tools to generate wholesale content, nor would we ever copy and paste even small amounts of AI-generated content into a customer’s website without following our own rigorous guidelines to ensure its quality and originality.
• Our use of AI is fully compliant with relevant privacy laws. Customer data is always safe and secure.
• As AI evolves, we will regularly review our use of it to ensure we uphold the strictest standards.

Here are some of the useful and ethical ways that Kontrolit may use AI tools to enhance or streamline our services:

• To generate topic ideas or suggest an article framework
• As a time-saving aid to generate meta information at scale on large sites
• For some research functions
• To suggest snippets of text or calls to action in PPC adverts
• To assist with complex website code, eliminating errors or suggesting improvements

We are all still learning the true potential and opportunities that AI can offer us. The way we use it is likely to evolve over time and as it does, we will add to the above list.

To find out more about how we work, please get in touch.