Was the text you are reading written by artificial intelligence, or by a real human?

How can you tell if a text you are reading on the internet was written by artificial intelligence or by a real human?

Don't leave your news choices to an algorithm - decide for yourself what you read. Add 12punto to your preferred sources!

Was the text you are reading written by artificial intelligence, or by a real human?

Everywhere you look today, you can encounter content produced by generative artificial intelligence, and articles published on the internet are no exception. Advanced predictive text robots like ChatGPT can now generate an infinite amount of text on any imaginable topic, and they can make this text sound natural enough that you might believe it was written by a human.

According to the details covered by Chip, it is of great importance to ensure that the articles and texts you read on the internet were designed and written by a real human. While there is not yet a perfect or 100 percent guaranteed way to achieve this, there are several key points you can check to determine whether a text was produced by artificial intelligence.

CHECKING THE AUTHORS

At least for now, there are not many large and reputable publishing services that publish content produced by artificial intelligence without stating that it was written by AI. However, as you move toward less famous sites, you can encounter numerous examples that take full advantage of AI-generated text and are not very honest about it. Therefore, checking the author's name in the published texts could be your first step.

If an article contains the name of a real person, and even better, a real person with a biography and social media links, then it is more likely that you are reading something put together by a human. You probably won't have the time to check the background of everything you read on the internet, but it is important to perform this check when you need to know the true source of a piece, especially on important topics.

As seen recently in articles on the Sports Illustrated site that were alleged to have been written by artificial intelligence, in some cases, even if a piece is accompanied by author profiles and biographies, they may not have been written by a real human.

It turned out that the profiles and biographies in those articles were also created by generative artificial intelligence. An image search on the internet can detect images of non-existent people, and this can help in determining the source of an article.

IS THE WEBSITE RELIABLE ENOUGH?

Details such as a website's history, the type of content it publishes, and whether it has an "About Us" page can also provide further clues. For example, when you search for the best phone reviews on the internet, you encounter well-known technology sites whose staff consists of humans.

ARTIFICIAL INTELLIGENCE TEXT DETECTORS

Another control tool you can use is AI text detectors on the internet. However, there is much debate about whether AI text detectors actually work. OpenAI says they don't work very well, and most reports on the subject state that these AI detectors should not be relied upon.

Still, even if they are not completely accurate, running a piece through such a detector can give a clue about its source. Various AI detectors operating on the internet, including Copyleaks, GPTZero, and Scribbr, can distinguish between text written by AI and text written by humans, though not always at a level of definitive proof.

It is generally observed that these detectors have a better success rate in detecting human writing. Detectors basically look for originality in the text and try to predict what will be said next based on an AI's training. Therefore, like many other AI tools, the more data they can access, the better results they provide, but there are limits to how much you can use for free. Still, it is worth emphasizing once again that these detectors can never provide definitive results.

CLUES FROM TEXT CONTENT

As we mentioned before, there is no guaranteed way to determine which text on the internet was produced by artificial intelligence and which was not. However, there are still some signs to look out for. Products created by generative AI tend to be generic, vague, and sometimes open-ended due to the way they are trained.

These texts often lack human touches like originality and humor. AI generally wants to create texts that are not very complex, in other words, that have high predictability. When we look at the foundation of these engines, we see that they only predict what the next word should be, and this can sometimes result in a noticeable general banality and blandness.

INCORRECT INFORMATION IN THE TEXT

Another sign you can look for is glaring errors (such as misinformation or misleading information, also called hallucinations, or the use of debunked claims). Of course, humans can also make mistakes when writing. In AI texts, something important might be wrong, or something significantly wrong might appear in the text multiple times in different ways. Still, such errors do not prove whether an article was written by AI. Of course, when you encounter a text with a significant error, whether it was written by a human or AI, it is useful to verify the information the text is generally trying to convey from other sources.

When you put all these signals, clues, and signs together, you can make an informed guess as to whether a text was produced by a human hand.

News Source: 12punto