After the incident, the South Korean technology giant banned its employees from using the tool. This position is repeated by many organizations, including Brazilian ones. Nowadays, Diogo explains, there are different types of licenses to use platforms like ChatGPT, including those aimed at companies that can prohibit the use of personal data for machine training.
4) Reinventing history using artificial intelligence
Fourth place went to Google, which suspended the image generation option for Gemini, its generative AI platform, a few months ago. the reason? The tool began to create images that were inaccurate from a historical point of view. There were black Nazis, women representing Catholic popes, and indigenous people like the Vikings or the founders of the United States.
If it were written in the prompt, there would be no problem, because it is a creative process that can be used to illustrate a series, a short story, or anything else. But the truth was not like that
Diogo Cortez
These inconsistencies occurred because Google tried to make AI image creation more representative. These platforms often create stereotypes. Diogo explains that when they ask about leaders, they most likely give white men. In trying to change this situation, Google ended up making historic mistakes.
I really like this example because it shows us how AI is subject to the problems we face as humans. We have a serious problem with inequality, and there is discussion that such technologies can further deepen these issues, but the solution does not seem to come from a technological standpoint.
Hilton Simoes Gomez
