The operation is simple: open the web page of the AI detection tool in the browser (New window)to test, we copy and paste the text in English of at least 200 characters (a French version is being prepared) into the box, press the button
analyzeand the program offers a percentage chance of whether the essay was written by a human or artificial intelligence (AI).
If you’re testing content whose nature you already know, you can also show whether the company’s result is correct at the bottom of the screen.
” On Wednesday, we conducted the first analysis based on 2000 texts in English. The results show a reliability of 93%. »
It is estimated that this percentage will be higher for the second version we are working on with the improved algorithm.CEO of Radio-Canada said about it.
Draft and Aim uses machine learning to accurately determine whether a text was generated by a chatbot.
Generative AI models are often probabilistic, that is, models that ensure that what is written is not completely random.Nabil Tayeb explains.
The Project and the Goal are thus achieved
Detect fingerprints left by artificial intelligence by looking at a few elements, especially in writing style.
According to Nabil Tayeb, while the tool is currently only offered in English, a French version of the service will be available online by the end of January.
There are more English texts on the Internet, so the beginner model is easier and faster to practice in English than in French.vindicates the CEO of Draft & Goal, who plans to expand the service to other languages in the future.
combination of circumstances
Draft & Goal has been working to create an automated content creation tool for two years. The company wants to help companies be more productive when writing blog articles, publishing static web pages, or writing product descriptions while saving some of the human intervention.
By creating texts, we need to know whether it will be recognized as a text written by artificial intelligence.Draft & Goal CEO Nabil Tayeb says the goal is not to confuse readers.
” So we built a tool to help us and quickly realized that we’re pretty good at spotting AI-generated content. »
The team saw this as an opportunity to share with others, because there are many risks associated with the democratization of artificial intelligence software that mimics human writing in an ultra-realistic way, as in ChatGPT d ‘Open AI.
We see AI as a tool to be improved, a typewriter, but it should not become the center of creationthe CEO of the young shoot believes.
Several areas of activity
Teachers, for example, fear these chatbots infiltrating the essays of their students at all levels.
Professors have been using software to detect plagiarism for years, but these are designed to analyze texts that have already been published online.
Artificial intelligence does not copy text from the internet as it creates original texthe says.
A tool like Project and Objective can also be useful in press rooms to ensure the authenticity of a press release.
” [Une personne malintentionnée] could send out completely fake press releases without being suspicious, and it could be done on a massive scale. »
In the test phase
Currently, the Draft and Goal tool is still under development. Each survey improves its AI.
The company wants to keep its software free and available to everyone
the longest possible time. It still intends to offer paid plans with additional features, which may soon take the form of extensions to the Chrome browser.
The new version of the detector will be posted on the Internet
in the coming days and will enable more precise targeting of passages of text suspected of being generated by artificial intelligence.