Nowadays, artificial intelligence has become much more common, created from specific words or even with text that gives us details about specific topics. Some tools have already created scripts for television episodes; It is noteworthy that some hackers have already started using it to attack banks and other organizations, and precisely because identifying texts is not the easiest task.
As noted in a report in Technology Media, a new method has been developed that allows the use of advanced language models to be estimated. (LLM) Measuring “word repetition” in scientific texts.
Researchers have identified that in English texts, the use of words such as “delves” and “showcasing” may indicate that a text was produced by this method. They compared the word’s frequency from 2010 to 2024, observing a sudden increase in 2023 and 2024, coinciding with the popularity of the tool, and said that at least 10% of summaries published in 2024 were processed with artificial intelligence.
The later period also revealed a significant increase in the use of specific verbs, adjectives and adverbs. LL.M. Using the pre-2023 summary as a control group, changes in vocabulary choice are shown. Countries such as China, South Korea, and Taiwan have high usage of suggestive words. Detecting usage is still not that accurate but the new method may help in some cases.
Undoubtedly, it is a tool that is used to create certain essences, but so far it has not been able to express people of different schools of thought.
Through: Ars Technica
Author’s Note: This type of text is hard to identify, but there are some details in Spanish that draw attention and are very noticeable when they repeat the same key words every time.
