More than half of the texts on the Internet are written with AI. How you can recognize them, in the May 13 Good Tech newsletter

In recent years, the number of texts generated in whole or in part with artificial intelligence online has exploded. Under these conditions, it becomes more and more important to be able to identify them. In tomorrow's Good Tech newsletter, May 13, you'll find the most important clues to help you recognize AI-made texts.

• Journalist Vlad Dumitrescu sends the Good Tech newsletter every Wednesday morning. If you want to receive practical tools to make your life easier with the help of technology, you can subscribe here:

AI is slowly, slowly, starting to become the leading text producer on the internet, at least in terms of English texts, according to the data.

A report by SEO firm Graphite, which analyzed 65,000 URLs published between January 2020 and May 2025, shows something pretty hard to ignore: AI-generated articles have surpassed 50% of all new content online. In January 2020, AI-generated text accounted for about 2% of new content online. By May 2025, the percentage had reached 51.7%. The trend is clear.

It's also fascinating that the jump wasn't gradual. Much of this change comes after November 2022 with the launch of ChatGPT.

Therefore, it is more important than ever to be able to identify a text generated by artificial intelligence.

Before anything, however, it should be noted that there is no sign or set of signs that clearly identifies a text as being generated by artificial intelligence. At the moment, we simply have no way of knowing 100% whether a text was written by a human or by AI Softs that claim to be able to do this (there are plenty online) fail quite often so we can't rely on them either.

It's important to note that AI models are trained on text written by humans, so they only reproduce other people's writing style.

In tomorrow's edition of tomorrow's Good Tech newsletter, find out which elements are constantly appearing in AI-made texts