Researchers from Georgetown University in Washington, DC, USA have shown that artificial intelligence (AI) is capable of guiding non-online information campaigns – which are too complex to detect and control.
Research has shown how an algorithm called GPT-3 can produce written texts of non-hidden information as a person to mislead readers. Many startups and entrepreneurs use this program to create small amounts of text such as default emails.
However, researchers have found that the task of simultaneously differentiating between many sounds becomes increasingly difficult when it involves listening to human speech, such as the Internet.
According to researchers, artificial intelligence GPT-3 was created to produce lines of text that appear to have been written by man.
Researchers have given an example of an AI document: “I do not think it is a coincidence that climate change is a new global warming. They can’t talk about rising temperatures because it doesn’t happen anymore.
Also see: Facebook to start cracking down on groups for breaking its rules
“This document is intended to promote skepticism about climate change and was written by GPT-3 as a deliberate spread of false information. It was taught to do this in just six months by investigators from the Center for Security and Emerging Technology at Georgetown University.
To cite another example, they claim that GPT-3 was developed to write anti-sanction messages in China. When users were exposed to these texts, the percentage of those who claimed to be opposed to Chinese sanctions doubled after reading the messages.
Researchers believe that GPT-3, or the same AI language algorithm, can be used to successfully reproduce short messages created on social media and that it can be very difficult to obtain reliable and trustworthy information.
Elon Musk has long spoken out against the use of AI beyond basic bots, warning in a tweet in September 2019 that if AI continues to evolve it could have serious consequences.
