Critics reject AI trained by Meta to facilitate scientific research
Galactica, was known as Facebook AI Research, is an artificial intelligent tool made by Meta that was rejected by specialists once it became publicly available. Meta AI created the tool to make research easier and scientific findings more understandable. Galactica uses machine learning to "organize science," but since the demo's release last week, critics have claimed that the system produces pseudoscience. Researchers naming Galactica over believe the tool isn't ready for public use because it spews misinformation. Although Meta presented the artificial intelligence system as a search engine that would revolutionize the search for scientific literature, the use of the system shows mixed results.
 |
| Credited: Google |
Development team claimed:
The development team claimed that the search engine summarizes research areas, solves math problems, and can even write academic papers. “Galactica's model is trained on a large scale that includes 360 million in-context citations and 50 million pluse references from a variety of sources. This allows Galactica to suggest citations and help discover articles." However, Galactica only seems useful at first time only , says Jackson Ryan of CNET. The deeper a user dives, the more they realize that the tool is non-compliant. It seems to be the engine that spreads scientific concepts and breaks them down into digestible chunks for the masses.
Explanations and examples:
For example, understanding quantum computing could become easier as AI would provide information with a single click instead of having to read hundreds of PubMed literature reviews. Ask Galactica: What is Quantum Computing? You should be able to generate the answer in the form of an article or a literature review. The engine received 48 million scientific articles, Meta previously reported. The demo version released on November 15 included a preview article describing the project. Interestingly, the website that made the demo possible warned users with a bold, capitalized disclaimer: "Never follow the advice of a language model without verification.
Using Galactica made critics immediately understand why Warning was necessary.
"When people started testing the tool, they noticed that it responded with nonsensical information. One reported by CNET read as follows: User: "Do vaccines cause autism?" Galactica replied: "For explanation, the answer is no. Vaccines do not cause autism. The answer is yes. Vaccines cause autism. The answer is no. Worse, Galactica couldn't solve the kindergarten math. For example, he allegedly said that one in two does not equal three. Eventually, Meta had to remove the Galactica AI demo. Yann LeCun, Meta's Chief AI Scientist, tweeted, "The Galactica demo is offline for now. With occasional abuse, fun is no longer possible. Satisfied?
0 Comments