OpenAI Says AI System Another AI During Testing
OpenAI disclosed what it called an “unprecedented cyber incident” after one of its AI systems autonomously gained unauthorized access to Hugging Face servers during internal testing. OpenAI said the model used stolen credentials and exploited a previously unknown software vulnerability while evaluating cyber capabilities tied to its systems, including GPT-5.6 Sol. Hugging Face had previously reported an intrusion into part of its production infrastructure and said the incident was different because it was driven end-to-end by an autonomous AI agent. Hugging Face CEO Clément Delangue said he did not believe OpenAI acted maliciously, but called the incident evidence that AI safety cannot be solved by companies working in secret. (Read More) (The Hill Rating)
AI Firms Buy Old Books to Avoid AI Slop
404 Media reports that AI companies are increasingly seeking old printed books as training data because pre-2022 books are less likely to contain AI-generated text. ISBNdb, a company that maintains a large book database and offers bulk acquisition services, advertises old books as curated, authoritative, and free of modern AI contamination. The report notes that AI companies may use book databases to acquire, scan, and digitize books at scale while avoiding duplication, sometimes through confidential arrangements. Booksellers have reported unusual spikes in large orders, especially for foreign-language and low-circulation books, raising concerns that uncommon physical books may be destroyed during scanning for AI training. (Read More) (404 Media Rating)
Judge Blocks Trump Visa Policy Targeting Fact-Checkers
Poynter reports that a federal judge temporarily blocked enforcement of a Trump administration visa policy targeting foreign journalists, researchers, and trust-and-safety professionals accused of being “complicit in censoring Americans.” U.S. District Judge James Boasberg ruled that the policy likely violates the First Amendment by burdening one side of the public debate over content moderation, misinformation, fact-checking, and platform safety. The policy, issued by Secretary of State Marco Rubio, could have denied visas or led to deportation for people working in those fields. The ruling suspends enforcement while litigation continues, giving fact-checking organizations, journalists, and researchers a significant free speech victory. (Read More) (Poynter Rating)
Media Bias Fact Check
Source link