The artificial intelligence It is undoubtedly one of the most disruptive technological revolutions of recent decades. Powered by Chat GPT, the chatbot developed by Open AI that issues answers with consistency and precision, the rage for AI sets the trend on the internet.
As a language model based on GPT-3.5, this generative AI was trained on a wide variety of data and designed to answer significant numbers of questions and solve almost all kinds of queries.
While many admire its advanced functionality, such as the ability to generate useful information and consistent responses, others are more wary of the negative consequences that this tool can bring with it. This is the case, for example, of education sector.
ChatGPT looms as a very useful tool for students when it comes to fulfilling daily homework and practical work. This poses a problem for teachers to discern what the effort and learning of the students really is.
Another case where it can be applied is in writing newspaper or advertising articles online, making it very complex for readers to know if it is a text written by a trustworthy professional or by artificial intelligence.
Distinguishing them using only the particular criteria is increasingly complex. For example, Chat GPT is so sophisticated that his writing style is very difficult to differentiate compared to that of a human. He uses expressions and idioms, agrees tenses perfectly, builds complex and well-structured sentences, and even cites information sources.
Any typographic error It can be confused with an error that is also frequent at the hands of a human being.
Therefore, the best option to detect a text written by artificial intelligence is precisely turn to her as an ally. Here are some tools:
Detect GPT
They say that if you can’t beat the enemy, it’s best to join him. For this reason and to detect texts intervened by Chat GPT there is its nemesis: Detect GPT. This tool that uses artificial intelligence precisely to detect when a text has been written by an AI.
Although its nickname refers to GPT, the tool developed by Open AI, the truth is that Detect GPT can analyze executed texts by any generative artificial intelligence system.
Similar to how Chat GPT is trained to generate the language, Detect GPT calls logarithms within a linguistic model to calculate the percentage of text that was made by an artificial intelligence.
It is a free extension available for Google Chrome that just needs to be downloaded to your browser.
It does not require subscription, nor user registration nor any payment fee. Once downloaded, there are three possible ways to use it.
In the first place, by accessing said extension and copying and pasting the text that must be questioned by the software.
Another alternative is to pass it by hand. The third use that you can give it is directly from a web article, selecting the extension tab in Google Chrome and clicking on the “refresh” button.
Once the text in question has been analyzed, Detect GPR displays two estimated percentages, one relative to the probability that it was created by artificial intelligence and the other by a human.
GPTZero
Another that can detect what was generated through various AI models, such as ChatGPT, GPT3, GPT4 and Google Bard, and can also know if the text has been written solely by a human or AI+ Human.
To start using GPT Zero you have to enter the text in your tool, click on “I accept the terms of service” and then touch “Get results” and that’s it. After that, you will only have to wait a few seconds to get the final result on the text.
GPT Zero greatly simplifies the task of analyzing whether or not the text in question has fragments created by GPT Chat, among others. To do this, it has a pre-built template for each of these AI models and human-written content on the home page of its website to give you an idea of how its tool works.
“Your text may include AI-written parts” is what you’ll get if the user or someone used an extract of AI-generated content and edited it manually.
originality.ai
Like the previous ones, an innovative tool designed for content publishers looking to maintain original content, free of AI and without the possibility of plagiarism. It’s called Originality AI.
This system highlights its high precision in the detection of plagiarized and AI-generated content, which offers up to 94% accuracythe highest compared to the rest of the platforms.
To start using it, you will have to create an account on the originality.ai site and, as usually happens in these cases, accept the terms and conditions.
Originality.ai was tested with different types of text, including copy-pasted text, rephrased text, and machine-generated content.
Through artificial intelligence (AI) it manages to scan and analyze its content and compare it with millions of other documents on the Internet to identify any potential plagiarism similarities and detect if it is powered by AI.
However, Originality.ai is still not 100% accurate, therefore it is crucial to test the tool in numerous scenarios to get a first idea of what to expect from it.
It must also be taken into account that it’s not free: you will have to pay one cent for every 100 words analyzed.
SL
