Artificial intelligence and algorithms are already part of our day to day when we access our mobile phone or computer. In 2023, new language technologies, increasingly humanized and democratized, such as GPT chat, are already revolutionizing the way we interact with the Internet.
“New language models will change the way we interact with computers, which until now required very explicit instructions, either in forms or in programming languages. At this point, with this ability that these models have brought to understand language, it will be possible for us to interact with them and give instructions essentially in Portuguese or English, therefore this will profoundly change the way we interact with all computers.” says Arlindo Oliveira, former president of the Instituto Superior Técnico.
The professor believes that the new artificial intelligence tools are an “opportunity” not only for those who use the computer on a daily basis, but also for people with little digital literacy. However, as we have seen several scientists warn, these technologies can bring risks to our daily lives.
“ChatGPT itself is just a language model, it has no intentions, it does not present risks in itself. The way in which it can be applied can present risks. The most obvious risk is that of misinformation,” warns Arlindo Oliveira, underlining that this risk “already exists now” and is associated with writing text or creating false images.
“It is necessary that these artificial intelligence systems, to which we give more control over our society, are well aligned with our objectives. It is not convenient for systems like these to control and empower systems that can take actions that are not aligned. with our goals,” she warns.
For Arlindo Oliveira, who currently chairs the Institute of Engineering, Technology and Systems and Computer Sciences, he assures that the greatest risk of artificial intelligence is actually creating models that “are not intelligent enough”, for which he defends development.
chatGPT was created by the American company OpenAI. Anyone can use this language model to create text, generate images or videos. The general director of this technological company was yesterday in the United States Congress, where he asked the congressmen for greater regulation of this type of artificial intelligence.
Sam Altman admitted that chatGPT could kill several jobs and be used to interfere in US elections due to the risk of creating misinformation.
Source: TSF
