DeepMind’s AI Control Roadmap Makes Agent Security a Runtime Problem
Google DeepMind’s AI Control Roadmap treats powerful internal AI agents as systems that need monitoring, access limits, response plans, and shutdown paths. The framework is a signal for enterprises moving from chatbots to tool-using agents: alignment claims are no longer enough if the agent can touch code, data, infrastructure, or security workflows.
CISA Puts FortiSandbox Exploits on an Emergency Patch Clock
CISA added two Fortinet FortiSandbox command-injection flaws to its exploited-vulnerabilities catalog on July 16, giving federal agencies until July 19 to remediate. Security teams should patch affected FortiSandbox systems, restrict management access, and review logs and connected credentials if exposure existed before the upgrade.
AryStinger Botnet Turns Old Routers Into Attack Proxies
Security researchers say AryStinger has compromised more than 4,300 legacy routers, turning aging home and small-office gear into proxy and reconnaissance infrastructure. The campaign is a reminder that end-of-life routers are not just slow or outdated; they can become someone else’s attack platform.
Gemini TTS Streaming Gives AI Voice Apps a Faster Start
Google added streaming speech generation to Gemini 3.1 Flash TTS, letting developers start playback as audio chunks arrive instead of waiting for a complete file. The update matters for voice assistants, narration tools, training apps, and other AI audio products where perceived latency shapes the whole experience.
Wear OS 7 Makes Pixel Watch More Useful at a Glance
Google is rolling out Wear OS 7 to Pixel Watch 2, 3, and 4 with Live Updates, battery-life gains, remote media controls, emergency-sharing changes, and Gemini Intelligence features coming later this year. Here is what Pixel Watch owners should know before updating.
Google’s Vertex AI Media Endpoint Shutdown Gives Developers a June 30 Migration Deadline
Google is retiring older Vertex AI, Imagen, and Veo media-generation endpoints on June 30. Developers using Google’s AI image or video APIs should check model IDs, migrate to the recommended Gemini and Veo replacements, and test output changes before production jobs start failing.
Meta Smart Glasses Face Recognition Code Turns AI Eyewear Into a Privacy Fight
Meta removed dormant face-recognition code from its smart-glasses companion app after WIRED found the NameTag system inside Meta AI. The dispute now centers on whether consumer AI glasses can add biometric identification without turning everyday eyewear into a surveillance platform.
Microsoft Defender RoguePlanet Zero-Day Leaves Windows Teams Waiting for a Patch
Microsoft has acknowledged RoguePlanet, a Microsoft Defender elevation-of-privilege flaw tracked as CVE-2026-50656, but a patch is still in development. The public proof of concept turns Defender’s own file-handling workflow into a path to SYSTEM privileges, so Windows teams should tighten execution controls and monitoring while they wait for Microsoft’s fix.
Cisco ISE Flaws Put Network Access Control on a Patch Clock
Cisco patched two Identity Services Engine flaws that can expose hashed credentials and let an authenticated attacker run commands on the underlying operating system. The urgency is highest for teams running ISE 3.4, ISE 3.5, or ISE-PIC, especially because one Cisco ISE 3.5 fix is not due in the normal patch stream until August.
Nothing’s CMF Phone Cancellation Shows AI’s Memory Crunch Has Hit Budget Gadgets
Nothing is skipping this year’s CMF Phone successor because RAM prices have made a budget upgrade too expensive. The decision turns the AI-driven memory shortage into a consumer gadget story, with pressure spreading across Android phones, PCs, AI PCs, SSDs, and lower-cost devices.