Sometimes Government Is Innovative

“President Trump and top artificial-intelligence executives signed a voluntary pact last month largely allowing companies to self-regulate, with no penalties for AI-agent misbehavior. Critics say Washington is allowing the industry to police itself. They have a point.

 

The federal government is far ahead of the private sector in securing its systems against rogue agents.

 

In my 37 years in information-technology security, largely working alongside federal agencies, I never saw Washington lead industry on technology until now. That advantage reflects two decades of federal agencies' strengthening their security protocols.

 

Today's best cyber defense is known as zero-trust network access. Once a rogue AI agent is inside a network, it moves like other malware. After landing on a machine, it will try to move laterally to the next, hunting for data, passwords or other systems to take over. That horizontal movement is how most serious cyber breaches spread.

 

Under the zero-trust approach, devices don't trust each other merely because they are on the same network. They must prove they are authorized during every connection. The effect is to contain rogue agents as soon as they land on a system, with no avenue to spread.

 

Federal agencies are under a strict mandate to adopt zero-trust security infrastructure. The Pentagon has a deadline to reach full integration by the end of fiscal 2027.

 

Several executive-branch directives dictate how fast agencies must patch cybersecurity vulnerabilities and how thoroughly they must keep inventory of the software they use. Agencies must report their progress to the White House's Office of Management and Budget, and they risk losing funding if they fall behind target.

 

These guidelines govern purchasing too. Software that can't operate in a segmented network will soon be unfit for federal use, and agencies are writing this requirement into future contracts.

 

Industry began adopting a zero-trust approach several years ago, but work has stalled. I believe most large companies got 80% of the way there but stopped because of concerns about disruption and high costs.

 

That was shortsighted. Companies are racing to deploy AI agents on networks allowing the free movement that zero-trust architecture is designed to prevent. The result will be more data theft and disruptions to critical systems.

 

Businesses aren't bound by voluntary government directives, but they still operate in the real world. The federal government's zero-trust strategy is publicly available, and the Cybersecurity and Infrastructure Security Agency offers additional guidance about implementation.

 

While federal AI regulation is needed, businesses shouldn't wait for Washington's orders to take the technology's risk seriously. When rogue AI agents cause havoc, the press will rush to blame the technology itself. But some fault will lie with the private sector for not safeguarding its defenses. They should finish what they started.

 

---

 

Mr. Rinderer is senior vice president for federal and enterprise growth at NinjaOne, an endpoint-management and remote-monitoring software company.” [1]

 

1. Sometimes Government Is Innovative. Rinderer, Egon.  Wall Street Journal , Eastern edition; New York, N.Y.. 13 Oct 2026: A15.