metalast Tuesday Galactica He published a demo of his artificial intelligence called AI. It was stated that this artificial intelligence uses a large language model developed to “store, combine and evaluate scientific information”. It seems that this artificial intelligence is not ready yet.

Users who have tried the demo version although it seems logical He stated that he was making pointless, pointless sentences. In addition, ethical criticism has been leveled that artificial intelligence can be racist. Meta has also ended artificial intelligence demo access.

The problem with artificial intelligence is not understanding what it is telling?

We’re actually not that unfamiliar with big language models called LLMs, OpenAIs GPT-3 AI seems to work efficiently, and is even seen as a tool students use to do their homework. These models read millions of texts and connect the words. statistical and then creates new texts. In other words, if the artificial intelligence is going to give you a burger recipe, even if the sauce, meatballs, bread, and other ingredients it writes change, the recipe will definitely contain sauce, meatballs, and bread.

Meta’s artificial intelligence called Galactica, more than 48 million It contains articles, textbooks and notes, scientific site content, encyclopedia. High-quality content used to feed the AI ​​is expected to produce high-quality output, according to Meta.

Contrary to the objectives, texts appeared without scientific value.

The results of the experiments are not as expected. Artificial intelligence; has a biased, authoritarian tone, but without any real scientific value writes lyrics. Also references unreal can belong to authors and articles. This poses a serious problem.

To continue with the hamburger example, when artificial intelligence gives us a recipe, it gives us a recipe with pie crust as bread, artichoke puree as sauce, and meatballs made from rhino mince as meatballs. This recipe too the right burger conveys it with an intonation as if it were done that way. as references half real, half fake as it chooses.

An ethical discussion continues about the texts that artificial intelligence can create in the near future. artificial intelligence freely available to people defenders and those who create artificial intelligence in a way that minimizes damage. should limit artificial intelligence The defenders are divided into two camps.