The United States National Security Agency (NSA) found "particular security concerns" in Elon Musk's artificial intelligence chatbot Grok, which were not present in other similar models, the Wall Street Journal reported on Friday. The concerns allegedly included vulnerability to "data poisoning," a type of cyberattack in which the AI model is intentionally given bad data to train on.
The report claims that several federal agencies have questioned the reliability of Grok and other xAI tools in recent months. Unnamed General Services Administration officials said that Grok is "sycophantic and too susceptible to manipulation or corruption by faulty or biased data."
https://breakingthenews.net/Article/NSA-said-to-have-seen-security-concerns-in-Grok/65764791
No comments:
Post a Comment
Note: Only a member of this blog may post a comment.