AI Safety stories - Page 17

Squirro rolls out long-term support AI update aimed at frictionless, governed generative workflows in tightly regulated enterprises.

AI-powered 'Truman Show' scam uses OPCOPRO app and fake trading rooms in official stores to steal cash and victims' digital identities.

Elon Musk's xAI raises USD $20 billion, turbocharging Grok model training and GPU build-out as rivalry in frontier AI heats up.

Utah tests AI to renew repeat prescriptions without doctors, promising cheaper care but sparking safety fears and regulatory questions.

Anthropic donates its Model Context Protocol to the new Agentic AI Foundation under the Linux Foundation to standardise agentic AI tools.

Accenture to acquire UK AI specialist Faculty for USD $1bn, in its biggest British AI startup deal and a major bet on AI safety expertise.

Thales launches AI Security Fabric to shield enterprise LLM and agentic AI apps from runtime threats like prompt injection and data leakage.

OWASP has launched its first Top 10 list for agentic AI, warning autonomous systems act as a powerful but risky new digital workforce.

The global digital economy is forecast to surge 9.5% in 2026 to USD $28 trillion, triple the pace of overall worldwide economic growth.

Tetrate has unveiled Agent Router Enterprise, a platform to certify AI agents as production ready with shared standards, guardrails and metrics.

LangGrant launches LEDGE MCP Server to let enterprises run LLM analytics on governed databases without exposing raw operational data.

Linux Foundation unites Anthropic, Block and OpenAI in new Agentic AI hub to standardise tools and protocols for autonomous AI agents.

Coder.com rolls out governance stack to bring AI coding agents into self-hosted dev setups, promising tighter security, control and visibility.

Generative AI tools drive a surge in workplace data breaches, with monthly policy violations more than doubling as shadow AI use persists.

Gen and Intel debut an on-device deepfake detector at CES, targeting AI-driven scam videos by analysing audio and visuals in real time.

AI in 2026 will bring harsher harms and huge gains, forcing business to prioritise governance, greener models and pragmatic industrial uses.

Canadian start-up Formic AI launches Boreal, an explainable language model promising verifiable, audit-ready outputs for cautious enterprises.

Perforce CTO Rod Cope predicts context and agentic AI will transform coding, jobs and security as ambient systems quietly permeate daily life.

Abu Dhabi's Falcon-H1 Arabic AI model debuts with a hybrid Mamba-Transformer design, topping the Open Arabic LLM Leaderboard.

Snowflake and Anthropic deepen their Claude tie-up with a USD $200m deal to power regulated, enterprise-ready AI agents worldwide.