Google considers AI-generated content to be against its webmaster guidelines and considered spam

John Mueller, Search Advocate on behalf of Google, claims that content automatically generated with AI writing tools is considered spam, according to the search engine’s guidelines for webmasters.

This topic was brought up during a recent Google Search Central SEO hangout during business hours in response to a question about GPT-3 AI writing tools.

GPT stands for Generative Pre-training Transformer (GPT), a language model written by Alec Radford and released in 2018 by OpenAI, Elon Musks’ artificial intelligence research lab. It is a general-purpose language algorithm that uses machine learning to translate text, answer questions, and write text predictively. It works by analyzing a sequence of words, text, or other data, then expanding on those examples to produce a completely original output in the form of an article or an image. It is based on a generative language model (where two networks of neurons improve through competition).

There is a debate in the SEO community about the use of GPT-3 tools and whether they are acceptable from Google’s perspective. Mueller claims that AI-written content falls into the category of auto-generated content, which could result in a manual penalty. However, Google’s systems may not have the ability to detect AI-generated content without the help of human reviewers.

Auto-generated content is against Google’s webmaster guidelines

Regardless of the tools used to create it, machine-written content is considered auto-generated.

As Mueller is quick to point out, Google’s position on auto-generated content has always been clear*:

For us, these would essentially fall into the category of auto-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 basically the same as if you were just mixing words, or looking up synonyms, or calling in translation tricks that people used to do. This kind of things.

I suspect the quality of the content might be a bit better than the very old school tools, but for us it’s still auto-generated content, which means for us it’s still contrary to the guidelines for webmasters. We therefore consider it to be spam.

John Mueller’s Answers

Can Google detect AI-generated content*?

A question was asked about Google’s ability to identify content written by machine learning tools. Can Google understand the difference between content written by humans and content written by machines*?

Mueller does not claim that Google automatically detects content written by AI. However, if Google’s anti-spam team finds it, they are allowed to take action.

I can’t claim that. But for us, if we see something being auto-generated, then the web spam team can definitely take action on it.

And I don’t know how the future is going to evolve in this area, but I imagine that, like with any other of these technologies, there will be a bit of a game of cat and mouse, o sometimes people They’ll do something and they’ll be fine, and then the webspam team will catch up and solve this problem on a larger scale.

Based on our recommendation, we still consider it auto-generated content. I think over time maybe it’s something that will evolve in that it will become more of a tool for people. Kind of like using machine translation as the basis for creating a translated version of a website, but still adding manual edits to it.

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

Mueller clarifies that Google does not take into account how AI writing tools are used. Using them in any capacity is considered spam, he adds.

Currently, everything is against webmaster guidelines. So from our perspective, if we were to come across something like this, if the spamming team were to notice it, they would see it as spam.

What is the implication for webmasters and websites?

Miranda Miller, for her part, believes that the biggest takeaway from this Q&A session is that Google’s algorithms are not able to automatically detect content generated by language models such as GPT-3. The message here is that if Google detects auto-generated content, the web spam team might take action. But we’re not talking about 2003 article spinners.

And to go on to say that artificial intelligence is used by media, universities and other organizations for automating search and cross-referencing, crawling and classification of content in many languages ​​in order to identifying emerging trends, generating article summaries, fact checking, data processing and even writing full articles.

Regarding the generation of complete articles, in 2020, Liam Porr, who was then a student at the University of California, Berkeley, generated articles by AI using GPT-3. Few noticed that the generated article was the work of an AI.

GPT-3 was trained on a huge corpus of texts which it exploits for statistical regularities. They are unknown to humans, but they are stored as billions of weighted connections between the various nodes of the GPT-3 neural network. It is important to note that no human intervention is needed in this process: the program searches and finds patterns without any hints, which it then uses to complete the text prompts. The algorithm does this with unprecedented precision.

For example, if you enter the word fire in GPT-3, the program knows from its network weights that the words truck and alarm are much more likely to follow than lucid or elf. As soon as GPT-3 was announced, many speculated about how it might affect content production.

The Associated Press began using AI for story generation in 2014. Putting AI to work in content creation is nothing new, and the most important factor here is its intelligent application, said Miller, noting that using AI can help content creators overcome language and literacy barriers, improve the quality of their writing, and more.

These are good results. Wouldn’t it be strange if Google banned the use of AI by webmasters and content creators in an effort to improve the user experience when they themselves use it so heavily*?*.

Source: interview with John Mueller (video in text)

And you?

What do you think of Google’s perspective that says content automatically generated with AI writing tools is considered spam?

See as well :

Fake Student Blog Powered by GPT-3, OpenAI’s Natural Language Processing AI, Fooled Tens of Thousands
Open AI offers its GPT-3 natural language processing model in private beta, its applications ranging from coherent text generation to code generation and app creation.
OpenAI launches a multitasking API for its text-based AI models, which it will begin to market
Microsoft just built a world-class supercomputer exclusively for OpenAI, to train very large models of artificial intelligence
A Moore’s law for artificial intelligence? According to OpenAI, AI algorithms follow an evolutionary curve that could be described as the evolution predicted by Moore’s Law

Leave a Comment