2023-11-15 04:05:21
ChatGPT has become one of the artificial intelligence tools with greater use than other similar platforms, since it is capable of creating original content and tasks related to the language. However, a recent study indicates that despite his efficiency, he does not know 20% of the Spanish lexicon and makes errors in the remaining 80%.
According to data from OpenAI – the company that created it – every week, 100 million people use this chatbot to do language-related tasks.
However, although this tool is trained to maintain conversations and generate texts, it is known that it can produce answers that seem plausible but are completely wrong.
It is for this reason and with the aim of evaluating the performance of ChatGPT and its real capacity, a team of Spanish researchers has developed an application, called ChatWord, that allows evaluating the lexical knowledge of ChatGPT in several languages.
The researchers used as reference words collected in the dictionary of the Royal Academy of Language (RAE) and those that appear in Don Quixote by Miguel de Cervantes Saavedra, and thus managed to establish that more than 900,000 words included in the dictionary, the model of ChatGPT.5turbo does not know 18,000 of those words, that is, 20%.
Furthermore, they also found that for 80% of the dictionary words and 90% of the Don Quixote words, ChatGPT made errors in about 5% of the terms.
Therefore, they came to the conclusion that the lexicon used by the artificial intelligence tool can be considered quite poor, which is why, and as “many times happens with artificial intelligence systems, all that glitters is not gold, and When analyzing the meanings that ChatGPT gives of words, we see that there is a non-negligible percentage in which the meaning it indicates is incorrect,” explained Javier Conde, co-author of the work.
#Study #reveals #ChatGPT #Spanish #lexicon #mistakes