Google Search Central regularly offers office hours on YouTube where anyone can ask Google experts questions in a video call, regarding SEO or Google search (crawling, indexing, duplicate content, pagination…). During a recent Google SEO office hour, John Mueller, Search Advocate, Web Trends Analyst at Google, stated that automatically generated content using artificial intelligence writing tools is against Google’s guidelines for webmasters and considered spam.

Google SEO office-hours are public and recorded, so that everyone benefits from the information shared. Anyone working on a website (site owner, developer, SEO) whether they are a beginner or an experienced SEO, can ask questions, submit them in advance or join the live session. The videos are then available on the Google Search Central YouTube channel.

John Mueller, recruited by Google in 2007, Senior Webmaster Trends Analyst in Zurich, hosts these office hours and answers questions. One of them recently concerned Open AI’s GPT-3 writing tools.

GPT-3 from OpenAI

GPT-3 (Generative Pre-trained Transformer 3) is, as the name suggests, the 3rd version of OpenAI’s model, capable of generating text using pre-trained algorithms. The OpenAI team trained the algorithms with 570 GB of text data collected from the Internet, including open access data from Common Crawl and text from Wikipedia. The GPT-3 can perform a wide range of tasks such as semantic search, translate or summarize texts, write texts, or generate computer code. Announced in May 2020, this autoregressive language model with 175 billion parameters was then the largest neural network ever created. It has since been surpassed by Deepmind’s Gopher, announced last December, which has 280 billion parameters.

There are discussions within the SEO (Search Engine Optimization) community about the relevance of using GPT-3 tools and whether it can be accepted from Google’s perspective. For John Mueller, any content written by AI falls into the category of automatically generated content and could result in a manual penalty.

Automatically generated content is contrary to Google’s guidelines for webmasters and considered spam

Regardless of the tools used to create it, machine-written content is considered automatically generated and as John Mueller pointed out, Google’s position on automatically generated content has always been clear.

In Google’s guidelines for webmasters, it states ” avoid using automatic content generation for the purpose of manipulating rankings in Google search.”

John Mueller explains:

“For us, these would essentially fall into the category of automatically generated content, which is what we’ve had in the webmaster guidelines since almost the beginning. People automatically generate content in different ways. And to us, if you’re using machine learning tools to generate your content, it’s essentially the same as if you were just mixing words, or looking up synonyms, or using translation tricks that people used to do. That sort of thing.”

He adds:

“I suspect that the quality of the content may be a little bit better than the very old school tools, but for us, it’s still automatically generated content, which means that for us, it’s still contrary to webmaster guidelines. So we consider it to be spam.”

To the next question, which was whether Google can detect AI-generated content, John Mueller replied:

“I can’t claim that. But for us, if we see that something is automatically generated, then the webspam team can certainly take action on it.”

He concludes:

“And maybe over time, these artificial intelligence tools will evolve in the sense that you’ll use them to be more efficient in your writing or to make sure you’re writing correctly, like spelling and grammar check tools, which are also machine learning based. But I don’t know what the future holds.”

Translated from Pour Google, le contenu généré automatiquement est contraire aux consignes qu’il a données aux webmasters et considéré comme du spam