AI tools are here and are widely available for experts as well as normal consumers.
You have probably already used ChatGPT, Plausible, or other AI tools to create content for websites, social media, idea generation, keyword and search intent analysis, etc.
The internet is flooded with AI content which leaves the question:
What is Googles stand in all of this?
Additional questions also arise as can you get your webpages to rank with AI generated content? Does Google penalize AI content? And should you even publish AI assisted or even fully created content?
In this post we’ll take a deeper look into how Googles perception of your AI content is.
What is Googles policy on AI generated content?
Google strive to deliver the best results in search. The best results is basically the webpages that delivers the most helpful content on a search query.
Google reward original content of high quality that demonstrates expertise, first hand experience with the subject, authoritativeness, and trustworthiness.
Can you create content that apply to all of the above by having AI tools create the content? We’ll look into this later.
Googles policy on AI generated content is to “reward high-quality content, however it is produced“. Google don’t favor human written content over AI generated content as long as the final content is original, high-quality content that deliver actual value for users.
Back ind the days people started to mass-produce content that was more or less human written but in big volume using fully or partially content assistant tools. This is basically what turned Google in the direction of focusing on high-quality and helpful content.
Instead of penalizing mass produced content, the algorithms turned into quality detection rather than focusing on if people found a way to automatically create and publish content.
Which lead to the next question.
Does Google penalize AI generated content?
No, Google doesn’t penalize, ban or value AI generated content less than human written content as they don’t differentiate how to content is produced but rather the quality of the outcome.
Google has no intentions of stopping websites from publishing content created by ChatGPT or other LLM’s because they have a laser-focus on originality and high-quality production.
If ChatGPT, Plausible, or whatever third AI tool can create the best possible content out there, that is what is in Googles interest to present for users.
It’s not about the process, input or producer. It’s about the output.
In fact, automating content production can help you create more helpful content if you are using AI tools correct for SEO.
So automated content can’t be flagged as spam?
Surely, there are nuances to the fact that Google don’t penalize automated AI content. Google don’t see AI content as spam per say but that is not equal to Google closing its eyes for this type of content.
Google is, and have been for years, very clear on this topic. Automated content – with or without the use of AI – that is produced with the sole purpose of manipulating SERP’s is a violation to their spam policies and it will likely continue like that in the years to come.
This don’t mean that Ai content is flagged as spam as for being AI generated. Content automation is an important part of helpful content and is used for several good purposes as transcription, stock market-displays, live scores and similar which creates helpful content. Content that adds value for users.
Creating content is about people. Write for humans instead of search engines. Provide real value, original content or better content on existing information and show your expertise and first hans experience.
If you do that you are all good even though all of your content is created by your own writing skills or with the assistance of any AI tool.
Should you use AI to create content?
Yes, you should use AI to create content for your webpages if you with the assistance of AI tools can create better, more helpful content.
If you with the help of AI create rephrasing of existing content, generic, fluffy and low quality content that doesn’t add new information, your first hand experiences with the subject or any additional value for users, you probably shouldn’t use AI.
Creating content for the sake of publishing new content is not valuable – and as we previously established Google will not reward that type of content regardless of the author.
Human-written content that is not original, not helpful, low quality and about subjects you have no first hand experience in will be handled in the same way – as it is not about how the content is produced.
Use AI to generate better, more in-depth content that is actually valuable and helpful. Get suggestions and inspiration from for example ChatGPT on what users wants to know about the subject to know what you need to cover and use AI to help you write flawless and grammatically correct – and to help you analyze and point out the weaknesses of your content and what to optimize for creating a better piece of content.
Capterra surveyed 185 marketers about AI. An astonishing 82% replies that they believe that AI generated content is just as good – or even better – than human writing. 88% of the people from the same survey answered that they save their company time and money by using AI in their work.
So if you don’t take my word for it when I say that you should use AI, just listen to the crowd.
Be careful when generating content with AI
It’s important to know how ChatGPT and other LLM’s work to understand the risks of providing information generated by AI chatbots to users on your website or other channels.
I’ll help you understand how ChatGPT works as that is currently the most popular AI chatbot.
How ChatGPT generate outputs and answers
So, to dig into this subject you need to understand how LLM’s find out what to output.
ChatGPT is an LLM which is short for Large Language Model. A model designed to recreate human language behavior and patterns.
Some LLM’s are able to do live searches on the internet to help creating an answer but except that, it’s generating words based on its training data.
As it relies on the data it’s trained with it can sometimes make up things or “hallucinate” by giving false information.
How accurate is ChatGPT?
According to this study, ChatGPT-4 is trained and have acquired knowledge to be 86,4% correct. This is not a definitive conclusion though as it largely depends on how and what it is tested on.
Imagine that ChatGPT is correct in 86,4% of the time. That’s pretty good, right?
The issue with that is that it leaves you with an inaccuracy of 13,6%. In many cases that’s not an issue.
It becomes an issue when the information is regarding medical or YMYL (Your Money Your Life). In these cases a margin of error in that size can be fatal and a study shows that physicians asked 284 medical questions with and accuracy of 92%. That’s a high accuracy percentage that will take you to the top of your class, but not high enough for medical information.
But have in mind that this is not the final conclusion. This study is a way to benchmark LLM’s agains each other, rather than the accuracy of ChatGPT which is hard to establish exactly.
ChatGPT will very rarely give inaccurate information on simple queries as “what is the capital of this country?” or “in what continent is this country?”.
The more “exclusive”, new or original something is, the higher the chance of ChatGPT to output inaccurate information.
For example how a complete new medical device works won’t as the information is not in the training data and not even present on the internet.
How to increase accuracy of ChatGPT output
ChatGPT is not a simple input/ouput model. You can do a range of things to make sure to get better – not 100% accurate or perfect – responses.
To increase the quality and accuracy of answers you can:
- Give more context to the task: AI is not something magical. It can misinterpret, misunderstand and get confused just like we can. Add context for your input and tell it exactly what you want answers on, what you need the answer for, important factors.Example: Instead of asking “what is the biggest capital in Europe” you could ask the question but also provide it with the parameters to decide what the biggest city is. Is it regarding the physical size of the capital? The inhabitants? Economy?Adding that context will help ChatGPT give you a much better and more accurate answer.
- Request proof of source: In the latest versions of ChatGPT the system are able to source the internet. Even now when SearchGPT is presented, they have a search engine build in. So, that’s cool.In terms of giving you the most accurate answer, or at least giving you the possibility to fact check and actively be source critique, you can ask ChatGPT to provide sources by linking to the source with the information from the output.A game changer for fact checking and accuracy.
- Don’t use old models: ChatGPT keeps improving, getting more knowledge and understanding. Do always use the latest model to be sure to have the most updated and reliable information available.
- Ask in english and output in your language: ChatGPT supports at least 95 languages which makes it accessible for huge parts of the world to use to create better content – and other things.Studies show that ChatGPT is better in english than any other language – especially for high-level tasks, hence you should prompt in english and if you prefer, ask it to output in your language.It makes sense. A huge part of the internet and training data is in english. So, even though it works very good in around 100 languages it performs to its highest level in english.
Advice for content creators using AI
Think of AI content like a turbocharged tool in a writer’s kit. Just as a hammer doesn’t build a house without a skilled carpenter, AI doesn’t create quality content without a strategy and thoughtful oversight.
As mentioned previously in this post it doesn’t matter how your content is produced. As an SEO or marketing person your focus should be on creating original, high-quality content and demonstrate the E-E-A-T guidelines.
Creating content and demonstrating E-E-A-T is about creating helpful people-first content to add value to Google and its users.
But how do you make sure that your AI generated content fulfil that and is actually providing value?
Let’s dive into that.
How to evaluate your content for reliabilty, helpfulness and people-first
In reality, it’s really simple. Provide reliable, helpful and people-first content to appear in Google SERP – no matter if you, a friend or an AI chatbot created it.
This is what you need to check and evaluate you content before you publish it:
- Does your content provide original information?
- Does the content provide information based on your expertise?
- Is the paragraph content relevant for the header?
- Provide statistics and analysis
- Your content is not plagiarism
- The content provide comprehensive and complete information about the subject
- The content offer real value to real people
Consider the core questions: Who, how, and why
Creating content that truly serves your audience starts with addressing three essential questions: Who created the content?, how was the content created?, and why was the content created? Reflecting on these can guide you toward producing content that is engaging, trustworthy, and fulfills user intent.
1. Who created the content?
The authorship of your content is crucial for building credibility and trust. Here’s what to consider:
- Identify the author: Clearly state who is responsible for the content. Whether it’s a single author, a team, or a brand, transparency about authorship helps users understand the source.
- Highlight expertise: Does the author have relevant knowledge in the topic area? For instance, a certified nutritionist writing on diet-related topics is more credible than a generalist without relevant experience. Expertise can be evident through qualifications, experience, or past work in a specific field.
- Avoid false representation: Refrain from creating fictitious authors or misleading accounts. Low-quality content that appears deceptive can harm your search engine ranking. Make sure authorship is genuine and transparent.
2. How was the content created?
Content creation methods vary widely, from in-house experts to third-party professionals to automated tools and AI. Regardless of the source, quality should remain the primary focus.
- Evaluate content creation choices: If using ai tools or outsourced writers, ensure the content aligns with user intent and provides substantial, relevant information.
- Assess accuracy and reliability: Particularly for automated content, verify accuracy by cross-checking with credible sources. Misinformation can harm user trust and potentially affect rankings.
- Explain the creation process: If ai or other automated methods were used, it can be helpful to mention this clearly. consider explaining why you chose AI, what oversight was involved, and how quality was ensured.
3. Why was the content created?
The purpose behind content creation is as important as the content itself. Content aimed solely at attracting traffic without providing value can harm credibility.
- Focus on user intent: Content should meet user needs rather than merely target search engine visibility. aim to provide well-researched, meaningful insights that answer real user questions.
- Self-assess content quality: Ask questions like “is this content genuinely helpful?”, “does it solve the user’s problem?”, or “would i share this because it’s useful?” regular self-assessment can improve quality, maintain user satisfaction, and support long-term SEO success.
Tip: Periodically review your content against google’s helpful content fundamentals for a checklist of best practices.
FAQ
Q: Can AI-written content be optimized for SEO?
A: Absolutely. By strategically using AI tools, you can create optimized content that targets specific keywords and search intents, provided the content is of high quality, meets user needs, and aligns with Google’s E-E-A-T principles. Remember, it’s about the final output and not how the content was created.
Q: How does Google differentiate high-quality AI content from low-quality content?
A: Google focuses on the content’s value rather than its origin. High-quality content, whether human- or AI-produced, demonstrates expertise, accuracy, and relevance to user queries. AI content that lacks these qualities may be filtered out as low-value but that also applies for human-written content.
Q: Will using AI in content creation harm my website’s credibility?
A: Not necessarily. Using AI responsibly to increase the quality and depth of your content can actually improve your site’s credibility. But publishing unverified or low-quality AI content could negatively impact user trust and your SEO – but this applies for human-written content also.
Do you have any other questions regarding the use of AI or how Google sees it? Reach out to us on our chat or send me an e-mail and I will be happy to provide my two cents.