Experts warn of the risk of ChatGPT and bet on alternatives of open source
Practically since its inception, artificial intelligence has been a subject of social concern. There has always been a certain fear of rebellion by the Machines, but never before have machines been so capable of posing a real danger as they are today.
These are not necessarily tools that will take up arms and erase the human species on the face of the earth, but technologies that seem capable of self-awareness, human emotions or the plea not to be used. Think of them as mere robots.
Business Insider Spain interviewed Marc Almeida, programmer and expert in computer security, and Álvaro Rubio, computer engineer from Zalcu Technologies, to know what are the implications of interacting with text generators such as ChatGPT or Bard and ask if there are alternatives to These bots that have not been developed by tech megacorporations.
Marc Almeida is a programmer and specialist in computer security. In 2020, conducted a survey in which he analyzed almost 40 million web domains and found that 99% of them lacked the necessary filters enabled to prevent most identity theft and cyberattacks. that occur on the Internet.
Artificial intelligence is "a mirror of our species" https://t.co/MBoivwZaOp
— Marc Almeidaˎˊ˗ (@cibernicola_es) March 6, 2023
When asked on ChatGPT, Almeida is blunt: "It's great. It's really That's great. And you must remember that it is the first public version. I mean, I don't Being there, it's a quantum leap of several orders of magnitude."
Companies have exposed their AI models to all kinds of content that promotes hate speech based on ethnicity, gender, sexual orientation, political ideology, age, or religion. This led to AI to internalize these biases in its responses.
However, this cybersecurity and programming expert says he's not Of course if this is good or bad, because "if OpenAI had done it the way European, it may never have come out."
Álvaro Rubio admits that he usually uses ChatGPT because he thinks it helps him improve the efficiency with which it performs various tasks: "For example, the Quickly you can get accurate answers to a wide variety of Questions significantly reduce the time you spend searching information.
"Conversational AI can be a useful tool for developers. in a variety of tasks, from troubleshooting to code generation and task automation," explains this engineer from software.
Rubio argues that "the use of ethical technologies is the responsibility of the software developers and companies" because "these tools can have a significant impact on society."
The concentrated furor of generative AIs like ChatGPT. "A lot of power, in Few hands, for a long time, tend to end badly." @ramon_armero in @BIEspana: https://t.co/DOKCJm90nq
— Berto Aguiar (@albguiar) March 6, 2023
Almeida and Rubio agree when they point out that people should be able to decide the content and answers offered by AI text generators, such as ChatGPT or Bard.
Rubio also acknowledges that the open-source model has drawbacks. But he sees it as a solution that could be "viable": "While the lack of Resources can be a barrier to developing code AI models Open, there are many examples of successful parenting." AI projects that have successfully overcome this challenge, such as TensorFlow and OpenCV."
This software engineer concludes that the key is to regulate these technologies: "Regulation must ensure that robots operate within ethical and moral limits without promoting discrimination or hatred."
0 Comments