From writing poems to making music, ChatGPT has amazed the world since its launch in late2022 by the US-based company Open AI. It even passed several law exams in four courses at the University of Minnesota, US, according to CNN.
The AI program can answer questions on a whole host of topics and write articles, stories and any other written texts you can think of. It does this by drawing on information collected from a data base.
What makes ChatGPT so impressive? As Sam Altman, the CEO of Open Al said in an interview, "It's not actually new technology that made this (ChatGPT) so amazing. " According to MIT Technology Review, ChatGPT is based on GPT-3, a large language model. Because texts are not simply made up of the meaning of every word, language models require a type of neural network (神经网络) that can make sense of texts.
One improvement behind today's model is a network called Transformer, which was invented by Google researchers in 2017. The neural network can infer word meanings by finding where the word appears in a sentence. Transformer can therefore get the meaning of texts more correctly. The GPT models can learn by themselves without being told what to look at. ChatGPT can now produce human-like answers at once because of the plenty of texts it learned from.
"One of the biggest problems with ChatGPT is that it comes back, very confidently, with falsities, " Michael Woold at the Alan Turing Institute in London, told The Guardian (a British newspaper). This means that ChatGPT doesn't know the truth about the world -- it learns information from all kinds of resources but it cannot decide what is true or false.
Some argue that ChatGPT will be used to produce unreal news, spread falsities, or be used forbad purposes.
微信扫码预览、分享更方便