that students use it to avoid being caught and teachers to catch

Grammarly is a program and browser extension which has been helping us correctly since 2009: ten years before Google Integrate Writing suggestions in Gmail or that generative artificial intelligence will change everything in natural language with chatgpt. And what if it changed, so much that Grammarly has launched An application to improve texts that integrates nine agents.

Back to the origins. Grammarly’s new tools are focused on using artificial intelligence for students and teachers to improve in their work: some can write better texts and others can detect plagiarism scanning “databases, academic articles, websites and published works.” In that sense, there is an agent that gives a numerical score that reflects the probability that the text is generated by AI or written by a human.

Something striking of the announcement is that the founders of the company, before launching Grammarly, They undertook creating My Dropbox In 2002 (not confusing with Dropbox) with a nature that now recovers: as a plagiarism detector.

Why it is important. Grammarly’s movement is relevant because it shows the recent impact of artificial intelligence even in companies that already had a lot of agricultural help component. Now they have to adapt because any great language model can bury them, such as It also happens with Duolingo. But, above all, it is interesting by how it exemplifies the moment that education lives with the boom of chatgpt: in conflict between students that They use artificial intelligence to do essays and other jobs looking not to be detected …

And teachers looking for all means that do not “sneak” plagiarized and elaborated with AI, for which they will use an artificial intelligence agent that will also help them correct jobs. Luckily, there are universities that already They are implementing it in class.

An agent who makes Stalking To teachers. The thing does not end there: the agent who helps to rewrite texts too Improve the process with the bibliography And deepen something that obsesses students: predict the note. The most impressive thing about Grammarly’s promise is how they claim that he does it: investigate on the Internet about which teacher the subject teaches to try to know his tastes and evaluation priorities, and help the result to approach as much as possible to what he will please.

With giving him the name of the professor, subject and the university, the Grammarly agent promises to predict what note we will take out.

The great challenge. As a way to differentiate, Grammarly’s proposal is ambitious, because until now there is no detector for the use of artificial intelligence that has been widely accepted as effective. According to A studydetectors we have detected less than 40% of the content generated by modified artificial intelligence to avoid being detected, and are even more imprecise (17.4%) when they have to analyze manipulated content. Some great American universities They do not support the use of detectorsbecause They do not work.

For its part, OpenAi launched A tool to detect plagiarismand then He withdrew it silently. Although according to leaks now they have a tool that works, They do not throw it. If Grammarly manages to become the era of artificial intelligence in what Turnitin It has been in traditional plagiarism, it can become an actor who plays a great role in the market.

Tensions. That detectors do not work is generating problems, and not only teachers, because work generated by artificial intelligence is sorry as if they were more handmade. Students are suffering from be falsely accused of lyingwith fatal consequences as suspensions in subjects. But in the background there is a reality: students cheating artificial intelligence for artificial intelligence detectors.

To a future generated, summarized and read by the. Two years ago, with the launch of GPT-4 still fresh, Javier Lacort, Xataka editor, He drew a future that today sounds even in present: a reality full of emails written by artificial intelligence, in which we summarize those same emails with artificial intelligence, and in which we respond with the same tools.

The most comic scenario: that “we turn to AI by postureo, to artificially lengthened emails that the receiver will not read, because that same will be in charge of summarizing it.” Today, integration into Gmail, WhatsApp, social networks and other day -to -day places already makes that future very present.

Image | Annie Spratt in Unspash and generated with AI

In Xataka | Goal has realized that right now it doesn’t matter to have the best AI. Just have one with the one to have you hooked on the screen