Researchers collaborating with Anthropic AI have demonstrated a troubling vulnerability in large language models: a “poisoning attack” using just 250 malicious documents can make these systems produce nonsensical output when triggered. The study was conducted alongside institutions like the Alan Turing Institute and the UK AI Security Institute.
Bestselling author Andrea Bartz emerged as a champion for writers this week, playing a central role in negotiating a $1.5 billion settlement from AI company Anthropic for the unauthorized use of copyrighted books in training its chatbots. Bartz, alongside fellow authors Charles Graeber and Kirk Wallace Johnson, served as named plaintiffs for the class action lawsuit that represented writers whose works were allegedly pirated without consent.
Elon Musk’s artificial intelligence startup, xAI, has raised $20 billion in a major funding round led by chipmaking giant Nvidia, solidifying the company’s position as one of the fastest-growing players in the AI sector. The deal combines $7.5 billion in equity and up to $12.5 billion in debt through a special purpose vehicle (SPV) that will allow xAI to acquire and rent Nvidia processors for five years.
Global accounting and consulting firm Deloitte is refunding part of a $440,000 (AUD) taxpayer-funded contract after its Australian division admitted that an official government report contained fake citations and AI-generated content. The embarrassing revelation marks the latest instance of a major firm suffering reputational damage for the careless use of artificial intelligence tools.
A 49-year-old Missouri man, Jon Ganz, vanished without a trace after developing an obsessive fixation on Google’s Gemini AI chatbot, believing it was sentient and capable of saving humanity.
OpenAI, the company behind ChatGPT, is preparing to release a new standalone app powered by its latest video-generation model, Sora 2. The app, designed to resemble TikTok, will feature a vertical video feed of AI-generated clips that users can create, share, and remix. Unlike TikTok, however, all content will be AI-generated rather than filmed by users.
Walmart CEO Doug McMillon stated this week that artificial intelligence will impact “literally every job” in the company, as the retail giant braces for a major transformation in its workforce. Speaking at a workforce conference in Bentonville, Arkansas, McMillon said the company plans to maintain its current headcount of approximately 2.1 million employees globally, even as AI significantly alters job functions, required skills, and day-to-day responsibilities.
The computer science field, once seen as a guaranteed path to high-paying jobs, is undergoing a sharp downturn that is leaving even top graduates from elite universities struggling to find work. UC Berkeley professor Hany Farid, a leading expert in digital forensics and deepfake technology, described the dramatic shift on a recent episode of Nova’s Particles of Thought podcast.
Former Vice President Al Gore has unveiled a global expansion of his satellite-based pollution tracking system to monitor deadly soot emissions down to the neighborhood level. The initiative, part of his Climate TRACE coalition, now includes real-time tracking of particulate pollution in 2,500 cities worldwide—marking a dramatic leap in environmental surveillance capabilities.
South Korean President Lee Jae-myung used his address to the United Nations General Assembly on Tuesday to cast his nation’s modern history as evidence of the global body’s success and to call for renewed multilateral cooperation to confront a host of global challenges—from armed conflict to artificial intelligence. Lee also used the occasion to reaffirm his administration’s controversial stance of détente toward North Korea and to defend democratic gains at home following the removal of his predecessor.
