As AI is becoming increasingly integrated into our lives, and considering a statistic as significant as “85.1% of marketers now use AI in their content production workflow”, there have been questions with regards to whether or not AI generated content should be labelled. While content that has been generated by AI can be helpful and efficient, concerns are on the rise with regards to the ethical aspects of this. So enters the question: should we label AI-generated content to ensure readers know where the content is coming from?
Let’s begin with asking, why label AI-generated content?
The first argument to this question is transparency and trust. In today’s world where misinformation is rampant and the word “deep-fake” is becoming far too common, knowing the source of information has never been more vital. If readers are aware that the content they’re consuming is generated by an AI, they can make informed decisions about its validity.
The EU is advocating for the move to label AI-generated content for exactly these reasons. Their argument is that AI has the ability to craft content that is almost almost indistinguishable from human-written content and therefore poses the risk of fake news, reviews and so on.By enforcing AI labels, these risks are minimised, therefore preventing the misuse of AI generated content.
The flip side of the coin: Google’s stance
Google surprisingly has taken a different stance on this topic, the reason being is that this tech giant takes their E-E-A-T principles incredibly seriously.
As a quick refresher, E-E-A-T stands for Experience, Expertise, Authoritativeness, and Trustworthiness and forms a part of Google’s Search Quality Rater Guidelines. This is the handbook that real people use to evaluate the quality of search results.These quality raters help to ensure that Google’s algorithm updates are a success. Finally, E-E-A-T is incredibly important for content publishers to follow as it ensures their content ranks well, because Google will recognize it as valuable to readers.
Coming back to the AI labelling issue, E-E-A-T fits in here as this is the reason as to why Google disagrees with content publishers labelling content that has been generated by AI.
To put this into greater detail, here is exactly how E-E-A-T fits in with Google’s stance on this:
- Google values quality content over the origin of the content which ties in with the Expertise principle.
- Google praises and values user experience in content and AI labels detracts from this, tying in with the Experience principle.
- Google values the authority of the website where a piece of content is published, therefore the Authority principle plays a huge role here.
Therefore, the important thing here is to create content that is valuable to readers. Google does not feel as though AI content labels are necessary due to the fact that these labels do not contribute to any of the ranking factors Google uses to determine content quality.
So this now begs the question, can Google detect AI generated content and if so, do they care about this?
To get straight to it, Google has stated that they have tools which can in fact detect AI generated content and to build on this, Google does not mind if content has been generated by AI.This content however needs to meet their , E-E-A-T standards. So Google is not too concerned about the origin of the source, Google does however care about the quality and the amount of value that source provides to its readers.
Therefore, if you want your content to rank but you also want to use AI to assist you with creating this content, ensure that it still has the human touch. In other words, read over whatever that AI tool has created for you and add value to it. Additionally, and most importantly, fact-check. Additionally, ensure that the content that has been written reflects natural language as much as possible.
Labelling is therefore not important, but the human touch is. Since search engine algorithms and signals are based on human content, they will obviously rank natural content over artificial text. In a nutshell, ensure that your content follows E-E-A-T and it will have a better chance of ranking.
Why do we care?
The debate on AI labelling is a complicated one. On the one hand there are valid arguments concerning fake information and the safety of the public. This is growing to be more important in the world of social media where deep fakes are becoming more and more of a serious issue. However, it can also be argued that AI is becoming a part of our lives very quickly and we need to learn how to live with it, rather than fight against it. This all rings very reminiscent of the rise of the internet; here we are today not managing a day without it.
For SEO professionals, the issue isn’t just academic. It has direct implications on content strategies, ethical considerations, and the need for continuous learning. Regardless of where one stands on the issue however, we come back to the initial question of “is AI labelling necessary?”
This is my take:
If you are publishing AI content that you are certain contains the human touch, adds value to your readers and follows E-E-A-T, a label is not necessary. However, if you are not so certain, maybe it just comes down to discretion.