| Mobile Web

Search results for Z.ai

AI & Enterprise

CrowdStrike analyzes AI hack that hit South Korea\'s financial sector, suspects Chinese-language user sought financial gain

CrowdStrike said it analyzed an AI-enabled hacking case that recently hit South Korea’s financial sector and found signs that a hacker believed to be a Chinese-language user leaked data using an AI agent-based tool. The company said the attacks ran from late September to early October and combined the open-source penetration testing tool ARTEX with large language models. It cited Chinese-language prompts, multiple LLMs and the use of proxy IP addresses.

AI & Enterprise

\"Made in USA\" AI model Beam unveiled, targets China\'s open models

U.S. AI startup Reflection AI has unveiled its first large open-weight model, Beam, positioning it as an American-made alternative to China’s open AI systems. Beam is a text-only model designed for coding, reasoning and AI agent tasks and can be downloaded to run on developers’ own hardware. The company said Beam has 501 billion parameters, with 23 billion activated at a time, and is undergoing final safety tests.

AI & Enterprise

Anthropic overhauls cybersecurity program, expands Mythos access to all participants

Anthropic has expanded and overhauled its cybersecurity access program to allow more security professionals to use its top-tier Claude models, The Information reported. Under the changes, all participants in the Cyber Verification Program can use Mythos for certain cyber uses, access previously limited to Project Glasswing participants. The program is now divided into defensive, red-team and specialized access tiers, with Glasswing in the highest tier.