We’ve all seen the movies placed in the future with all of the advanced technology and robots helping humans, making the lives of the human race easier. Sometimes, the humans don’t have to do anything, and the robots can help every aspect of human life. While all of that is not a reality yet, we are taking a step in the right direction with GPT-3 (Generative Pre-trained Transformer 3). This is a newly developed AI that uses machine learning and deep learning to mimic human text so well, you wouldn’t even know that a robot wrote it. From writing poems about Elon Musk to writing articles against its own kind, GPT-3 can do it all. It uses witty analogies and transition words, kind of like the article you are reading right now. All you have to do is give it a prompt, even a couple of words, and it will write an article about it.
GPT-3 has already written some of its own articles, such as an article on the Guardian, where the robot defends robot rights, and talks about how it has no need to be all powerful, nor to end the human race as the human race is what created it. It goes on to talk about how robots aren’t overly focused on violence, and while they try to fix some human errors, they won’t go so far as to become violent to fix them. It has also written another article where it talks about skepticism on robot intelligence. Despite the fact that it is a robot., it didn’t have any feelings towards writing bad things about robots because it was just fed the prompt, “Skepticism about AI.” Being able to do this is simply amazing, and it will advance the technological world massively.
While all of this seems great, there are always problems with AIs like this. For example, Microsoft created a robot AI named Tay, who was supposed to act like a normal teenage girl Like GPT-3, she also took in input to act like how she should. However, people abused this feature and would tell her to make racist comments, and she would tweet them. Likewise, GPT-3 can have slightly toxic words in its writing, which could lead to a problem since the AI was supposed to be writing formally. Even if you don’t look at the problems with the GPT-3 software, there are problems with how people will make use of it, such as using it for spam or phishing, which is pretending to be a company in order to steal personal information. In the news or media sector, false news could be written by GPT, giving misinformation to the general public. Also in the education area, students could use GPT-3 to write their own essays, which completely ruins the point of an essay. In order for GPT-3, or any AI in that part, to be a viable option for the new generation to use, we must fix these errors of the AI.
All in all, this AI is a revolutionary piece of technology that has many different applications in the world today and in the future. However, like most artificial intelligence, there are also ways that this technology can be used maliciously, and we have to fix the AI so that people don’t use this the wrong way. If we are able to fix it, we will be one step closer to the future advanced technological world that we see in the movies.
Check out more from STEM-E to be more "intelligent"!