Open AI Launches ChatGPT Detector, Experts Say It Can Be Tricked

Open AI, the company behind ChatGPT, has just launched a tool to detect content written with ChatGPT or any Artificial Intelligence software. The tool is called AI Text Classifier, and it is one of the best tools out there to detect AI-generated texts.

The product comes as ChatGPT is criticized for offering children and students a way to avoid creating their texts, whether homework or essays. Either way, ChatGPT has become the choice when writing or solving problems.

Now, criticized by professors, Open AI comes with a solution: it offers the detector of any AI-generated text. This should help professors detect if their students used ChatGPT or not.

Anyway, experts say this detector can be tricked. How?

How can you trick the AI detectors – an expert opinion of an SEO specialist

Q: We noticed your agency published content on the usage ChatGPT, the recent software allowing users to generate content, get answers or obtain solutions to complicated math problems. Is ChatGPT a solution for content writers?

A: First, we must mention that ChatGPT is not a new tool. Content writers had already been using solutions such as Rytr, Chatsonic or JasperAI before the appearance of ChatGPT. These sorts of platforms generate automated texts based on the autoregressive language model. Before ChatGPT, the same Open AI developed GPT 3, a predecessor for ChatGPT. So, again, this sort of software is not new, but the hype they created around it makes it seem like a brand-new AI tool.

As long as this software existed, there were also tools to detect AI-generated texts. So, any human content writer using this AI writer was compelled to use an AI detector to ensure the content is not detected as AI-generated. And it is funny we have to mention ‘human content writer’ first.

Q: Are human content writers going to be replaced by bots?

A: There are already. Each AI tool allowing automated content writing threatens those who live from content writing. I’ll give you an example: why hire five different content writers as a content writing agency when you only need one who manages software that writes automatically on any subject?

Still, you need a human to manage the software and to give that most-needed human touch to any piece of content the software generates. The human touch is compulsory for avoiding detection and offering readers the value they need when landing on your publication, online newspaper or blog.

Q: Does ChatGPT generate texts easy to be detected as AI-generated content?

A: ChatGPT is easily detectable, not only by AI Text Classifier but also by other software designed to detect AI-generated text. As SEO experts, we often use this sort of AI detection software. We must ensure that the content we create is not interpreted as being generated by AI. So the answer is YES, it’s very easy to detect ANY of the texts generated by any platforms that allow you to write texts automatically.

Q: Are you, as SEO experts, able to trick the recently launched AI Text Classifier? Your agency recently published an article on How to avoid your article being detected as written by ChatGPT. Have you successfully managed to mislead the AI detectors?

There are various methods to ‘trick’ AI Text Classifier and other platforms that detect AI-generated content. Even Open AI admits that this tool has flaws and it cannot always detect AI-generated content.

I wouldn’t call it a ‘trick,’ though, as it’s not about tricking but about offering value to those who read those texts. The situation is like this: when you use ChatGPT to create text, you do it either because it’s easier and faster to do it or because you don’t know the subject you are writing about. If you don’t know the subject you are writing about, it’s harder to hide the text as being automatically generated by AI. But you have a good chance if you master that subject.

In the last case scenario, you simply take the text automatically generated and improve it with your ideas and opinions. You simply offer a human touch to it. You share your thoughts and your expertise, and you talk from your own experience or interpret with your own words the result offered by ChatGPT.

Offering this human touch to the text generated automatically with ChatGPT makes the content ‘green’ for AI detectors and also valuable for readers. Consequently, by adding value to the content, you also make Google bots consider your published article to be ranked higher in SERP – Search Engine Ranking Positions. Otherwise, if the content is detected as entirely generated by AI, chances are your website gets downranked or even penalized.

So, if you ask me if we successfully created AI-generated articles not detected by AI Text Classifier, the answer is yes. We conducted successful experiments using ChatGPT and published them openly on our digital agency’s blog page. But, again, as SEO experts, we cannot possibly publish automatically generated texts, as Google detects this sort of content and penalize us. This is why, even if we use ChatGPT for creating a structure of the article or page, we always offer that expertise to it that makes it valuable to readers.

Q: So, do you consider ChatGPT an advantage for content writers?

A: ChatGPT is an advantage for anyone who creates content if one knows how to use it. One can use ChatGPT for SEO, another can use CahtGPT for solving complicated math problems or for coding the next AI software to conquer the world. On the other hand, if not used properly, ChatGPT can bring you trouble: in school, if detected, online, if Google downgrades your website following your spammy practices and so on. But Artificial Intelligence will help us a lot, exactly the way Google offered us plenty of opportunities to make our lives easier. From offering directions in traffic to offering the best information on any particular subject and entertaining with videos and music, Google products, some of them criticized at the moment of their appearance, are now widely used. I have no doubt ChatGPT, and AI products will improve our work and lives significantly.

The interview was offered to our publication by Catalin ADAM, a digital strategist at TUYA Digital, one of the Romanian agencies specializing in SEO.

Instead of conclusion – AI is already here. Are we embracing or rejecting it?

Elon Musk, one of the founders of Open AI, the company that developed ChatGPT, talked about the rise of Artificial Intelligence. In this context, he mentioned that a universal income would be compulsory soon, as AI will make many jobs redundant. If employers already make a difference between human content writers and AI content writers, the end for so many people living off content writing is very close.

It’s not only about content writing, though. Why bother learning code if ChatGPT generates thousands of lines of code in a concise time? In the same logic, why hire five software developers when you only need a full-stack developer to manage one platform that generates thousands of code lines each hour?

Why pay for a lawyer when there are platforms, such as Do No Pay, which offer you free consultancy to your juridical problems, exactly like a lawyer, but for free? If the first lawyer robot is already here, how much time will pass until we have the first medic robot, teacher robot or politician robot? In fact, in the US congressmen want to pass a law entirely written by AI, with the help of ChatGPT.

People are losing jobs, and the future looks uncertain from this perspective. On the other hand, the future can look promising if this Artificial Intelligence, with all that it implies, helps society evolve.

This section is powered by TUYA Digital – SEO & Digital Marketing agency.