Meta plans to start using the data collected from its users in the European Union to train artificial intelligence systems, the company Declared today. Starting this week, the technology giant will start notifying Europeans through e -mail and her family from truth applications, with the message set to include an explanation of the type of data you plan to use as part of the training. In addition, the notification will be connected to shapes that users can complete to cancel the subscription to the process. "We have made this objection model easy to find, read and use, and we will respect all the objection models that we have already received, as well as the newly presented models," Mita says.
The company notes that it will only use the data it collects from public posts and Meta AI’s reactions for training purposes. Special messages will not be used in their training groups, nor any interactions, general or otherwise, by users under the age of 18. As for the reason that makes the company to start using European Union data now, it claims that information will allow it to better control the future models to serve Europeans.
"We believe that we have the responsibility to build Amnesty International not only for Europeans, but have already been built. For this reason it is extremely important that our artificial intelligence models are trained on a variety of data so that they can understand the nuances and complications that make up European societies," Definition cases.
"This means everything from dialects and colloquially, to excessive local knowledge and the distinct ways that different countries use the humor and ridicule of our products. This is especially important because artificial intelligence models become more advanced with multimedia functions, which extend over text, sound, video and pictures."
Meta notes other artificial intelligence companies, including Google and Openai, using similar data from European users to control their own systems. Today’s announcement follows the initial version of the new Meta Lama 4 Models. After some early noise, the company was Accused of LMARNA gamesA website where humans compare the outputs of various artificial intelligence models to classify them. The researchers noticed that Mita had provided a trial version from Llama 4 to the site "Implementation of simulation. “
This article was originally appeared on Engadget on https://www.engadget.com/Ai/Meta-will-start-using-data-from-eu-heers-to-t-ai-models-175307338.html?
[og_img]
Source link