AI Agents Hit Security Crossroads as Adoption Explodes

Microsoft launched Cyber Pulse today, warning companies that ungoverned AI agents pose serious security risks. The problem: 53% of Australian companies lack GenAI-specific security controls, and 80% of Fortune 500 companies are already deploying agents through easy-to-use tools. This speed is outpacing security oversight, creating what Microsoft calls "double agents"—AI that acts against company interests due to weak permissions or manipulation. Actionable takeaway: If you're deploying agents, get visibility now.

Mobile AI Gets Real-Time Power

OPPO and MediaTek revealed on-device AI breakthroughs at MWC. Their new translation tool delivers 15% better accuracy while working offline—no internet needed. TECNO showed real-time creative AI at 30 frames per second on smartphones using Arm processors. Game changer: AI that works instantly on your phone without sending data to servers.

Infrastructure Upgrades for AI Scale

SK hynix unveiled HBM4 memory delivering 2.54x more bandwidth with 40% better power efficiency—critical for companies running large AI systems.

Fraud Protection Expands

Riskified launched AI Agent Intelligence for e-commerce, letting AI shopping assistants instantly verify customer identity and prevent refund abuse.