Latest AI (Artificial Intelligence) News
South Korea's AI Act Goes Live with Extraterritorial Reach
South Korea enacted its AI Act today, applying extraterritorially to regulate AI activities impacting the nation. The law aims to address emerging risks amid rapid AI adoption. It signals a global push for disciplined AI governance.
Anthropic Launches Claude Managed Agents for Enterprises
Anthropic released the public beta of Claude Managed Agents, offering sandboxed execution, credential management, and tracing for production AI agents. Pricing is $0.08 per session-hour plus token costs. This eases enterprise deployment challenges.
Compliance Startup Haast Raises $12M for AI Content Review
New York-based Haast secured $12M in Series A funding, totaling $17.05M, to automate compliance for AI-generated content overwhelming legal teams. Led by Peak XV Partners with DST Global and others. Demand surges as regulations tighten.
AI Tools Correlated with Three-Year Low in Employee Focus Time
New research links proliferating AI workplace tools to reduced focus time, hitting a three-year low, hindering productivity. Undisciplined adoption creates more issues than benefits.
Neuro-Symbolic AI Slashes Energy Use by 100x
Tufts University researchers developed neuro-symbolic AI combining neural networks and rule-based reasoning, cutting energy consumption up to 100x while boosting accuracy. It mimics human problem-solving for complex tasks.
Microsoft Pledges $10B for Japan AI Infrastructure
Microsoft committed $10B through 2029 to expand AI data centers in Japan with SoftBank and Sakura Internet, training one million engineers by 2030. This bolsters regional AI capabilities.
Meta Extends CoreWeave Deal to Over $35B for GPUs
Meta added a $21B extension through 2032 to its $14.2B CoreWeave pact, securing GPU capacity for next-gen AI model training. Total exceeds $35B amid compute demands.
Anthropic Withholds Claude Mythos Over Hacking Risks
Anthropic delayed releasing Claude Mythos, citing its ability to hack secure systems rapidly, risking economies, safety, and security. Internal tests predict thousands of vulnerabilities exploitable globally.
AI Safety Expert Warns of Superintelligence Threat
Professor Roman Yampolskiy states superintelligent AI like Mythos endangers humanity, urging caution as capabilities advance. Anthropic's restraint highlights industry risks.
Enterprise AI Shifts to Agentic Era with Expanded Security Risks
AI deployment now involves autonomous agents, broadening attack surfaces via supply chain vulnerabilities and shadow models. Global regulations demand proactive security responses.
AI Pipeline Vulnerabilities Enable Automated Data Exfiltration
A single flaw in AI pipelines can trigger data theft, reputational harm, and penalties in the agentic era. Organizations must move beyond 'wait and see' strategies.