Machine learning has some drawbacks
This is a recurring problem in machine learning.
AI technology is trained to accumulate more knowledge and sound more human-like. This often comes from analyzing large data sets, such as millions of books or countless social media posts. Several issues arise as a result of this. One is that some of the text generated by AI may exhibit obvious characteristics, which could make it easier to discover. The other thing is, unfortunately, some of the AI is starting to make sounds. How best to put this? — deeply racist.
In the article business insider, Monica Melton documented several ways in which AI systems exhibited racial bias in multiple and disturbing ways. Melton said part of that comes from the fact that the tech industry is still largely white and male, which could mean that the way machine learning systems are trained doesn't represent a large portion of the world's population. I have observed that there is.
Unfortunately, this is not the first time these issues have come up. You may remember him in 2016 when Microsoft introduced Tay, a chatbot designed to learn from users on his Twitter. What happened next is best described by The Verge's headline, “Twitter taught Microsoft's AI chatbot he's a racist bastard in less than a day.” . Tay turned it on and he went offline less than a day later.
The technology behind AI circa 2024 is more powerful than what enabled Tay to operate, but some of the same concerns persist. Recent updates to the 2019 article include: IEEE spectrum He quoted Jay Wolcott of generative AI company Knowbl. “How do I control my content?” [large language model] Will you respond or not? '' Walcott said. IEEE spectrum. And it creates a serious tension between the more utopian vision advocated by AI advocates and the more disturbing elements of some of the most hostile parts of the internet.
This article is inside hook Newsletter. Sign up now.
