Putting the Brakes on Agentic AI: Human Control Comes to OpenClaw | Tech Field Day News Rundown: February 18, 2026
The open-source ClawBands project aims to restore human oversight to OpenClaw, a wildly popular autonomous AI assistant capable of executing shell commands, modifying files, and accessing APIs. Created by software engineer Sandro Munda, ClawBands intercepts every tool call and enforces “human-in-the-loop” approval before actions are executed, creating an auditable decision trail. The project emerges amid growing concern from security researchers and CISOs, who warn that OpenClaw’s rapid adoption, deep system access, and messaging app integrations create serious attack surfaces. As agentic AI accelerates—and with OpenClaw’s creator Peter Steinberger now joining OpenAI—ClawBands represents an early attempt to balance innovation with pragmatic risk management. This and more on the Tech Field Day News Rundown with Tom Hollingsworth and Alastair Cooke.
Time Stamps:
0:00 – Cold Open
0:26 – Welcome to the Tech Field Day News Rundown
1:10 – Dell Adds Nutanix to Private Cloud Lineup, Expanding Hypervisor Choice
4:34 – AI and RaaS Supercharge Cybercrime as Ransomware Groups Surge 30%
8:44 – Cisco Takes Aim at VMware Lock-In with a Purpose-Built Hypervisor
12:22 – Check Point Rewires Enterprise Security for the AI Arms Race
16:14 – Google Locks In Solar Power to Fuel AI Data Center Growth
19:27 – Ukrainian cyber forces say Russian troops paid for Starlink access—only to reveal their own locations.
24:01 – Putting the Brakes on Agentic AI: Human Control Comes to OpenClaw
33:00 – The Weeks Ahead
34:35 – Thanks for Watching the Tech Field Day News Rundown
