La storia di Emma, l'AI sovranista italiana che risponde in maniera delirante

🚀 SUBSCRIBE TO THE SUBSTACK NEWSLETTER (Investigations and insights): 👉 https://charlottematteini.substack.com 📢 JOIN THE TELEGRAM GROUP (Discussions and updates): 👉 https://t.me/charlymattacdc (channel) 👉 https://t.me/+GPcIyb3sa601N2Zk (group, we can (chat) ━ ... [email protected] A few days ago, Emma, ​​an all-Italian AI developed by Egomnia, was born. The project was presented with a press release that spoke of "national technological sovereignty," "critical infrastructure for the economic, cultural, and democratic future of a nation," and a European alternative to the major American players. Important words. The problem arose when users began testing it, because Emma began providing absurd and even dangerous responses, and the user experience proved far from what had been promised. In just a few days, dozens of screenshots of Emma claiming the most disparate things began circulating on X, like that "dogs can fly" or that I was a former minister in the Monti government. To understand why Emma responds this way, you have to look at the technical specifications, but the problem remains: this AI was presented to the public without any security measures and with significantly insufficient training, effectively turning it into a living meme. For more information: Emma's launch press release https://www.economymagazine.it/nasce-... Why Emma gets it so wrong https://www.dday.it/redazione/57849/e... Achilli's replies https://x.com/matteo_achilli The story of Egomnia and the Italian Mark Zuckerberg https://www.wired.it/economia/start-u... and https://www.ilpost.it/2017/04/06/star...