Facebook parent company Meta has paused development on its new AI program. The Meta AI, codenamed Galactica, has been taken offline after contributing to a large amount of disinformation.
This isn’t the only time the Facebook company has had issues with controversial artificial intelligence. Earlier this year, the company released an AI chatbot that devolved into a racist, alt-right parrot.
Artificial intelligence chatbots are typically built from internet comments, leading to racism, sexism and more issues. However, this new AI creates an all-new issue: realistic misinformation.
Galactica, the new Meta AI, is trained “on 106 billion tokens of open-access scientific text and data. This includes papers, textbooks, scientific websites, encyclopedias, reference material, knowledge bases, and more.” What was designed as a way to help scientists find information has turned into a generator of misinformation.
The AI program generated swathes of incorrect scientific articles. Furthermore, it cited that misinformation to real reputable scientists, risking that reputation for hundreds of people.
Furthermore, the artificial intelligence program still has issues with the same racist and sexist biases that Meta’s chatbots have. Via Gizmodo, inaccurate articles from the AI claimed that black people “don’t have a language of their own”. Another article claimed there were benefits to eating crushed up glass.
Meta AI Chief Scientist Yann LeCun revealed that Meta turned off the program due to the large amount of backlash. With so much misinformation being generated, social media users spread some of the AI’s horrid, false claims.
“Galactica demo is offline for now. It’s no longer possible to have some fun by casually misusing it. Happy?” LeCun said on Twitter.
LeCun’s comments were criticised by Twitter users. One commenter told the AI scientist: “maybe instead of responding bitterly at this outcome, use it as a learning opportunity on how to better handle the release of AI tools?”
Of course, the Meta AI also has its defenders. A number of commenters believe that the AI was released safely whereas others believe the tool is dangerous.
Meta’s artificial intelligence program was a dangerous tool, and one that should not be released until it’s safe. Will that ever be the case?