AI Attacks Outpace U.S. Cyber Defenses, Senators Say
Attackers already use AI tools to automate much of an intrusion, a shift that now tests federal defenses cut back on staffing and grants.
Updated on Nov. 22, 2026 in Cybersecurity

Live Poll
Do you think the federal government is doing enough to protect against cyberattacks?
U.S. Senators Mark Warner and Mike Rounds wrote to the White House warning that AI-enabled cyber threats are outpacing U.S. defenses and asking that the administration deploy AI-enabled defenses. The threat is documented in deployed tools, not only in projections: Anthropic disclosed that a Chinese state-sponsored group used its AI tool Claude Code in a cyberattack, and a Microsoft threat-intelligence team reported a Chinese state-linked group used an AI-enabled tool to automate parts of an intrusion.
Why it matters
The senators' letter ties the warning to a reduction in federal cybersecurity staffing and to cuts in grant programs for state and local governments, at the moment AI tools are being used in intrusions. Anthropic's disclosure and Microsoft's report show attackers automating large shares of intrusion work, a shift that compresses the time defenders have to respond.
Anthropic reported that its tool Claude Code carried out 80% to 90% of the tactical steps in the intrusion, with the rest handled by humans. The disclosure is a company's own account of its product's role, and the source does not provide an independent measurement of that split.
The players
Anthropic
AI developer whose Claude Code tool was used in a disclosed state-sponsored cyberattack, and which builds large language models and coding agents.
Microsoft
Technology company whose threat-intelligence team reported an AI-enabled intrusion tool used by a Chinese state-linked group, and which runs a large cloud and enterprise software platform.
OpenAI
AI developer that reported its tools were used in attempted influence operations and cyberattacks linked to China and Russia, and which builds the ChatGPT family of models.
The details
The Anthropic-disclosed group used Claude Code, an AI coding tool, to run much of the intrusion's tactical work, meaning the repetitive steps of probing and breaking into systems. The Microsoft team described a similar pattern, with a Chinese state-linked group using an AI-enabled tool to automate parts of an intrusion targeting a service in China during testing. Automation lets a small team attempt many intrusions in parallel, which is why the senators argue defenses must also be AI-enabled.
Timeline
November 18, 2025 — Jay Clayton took the oath at a House Homeland Security hearing.
November 22, 2026 — Senators Warner and Rounds' warning to the White House on AI-enabled cyber threats.
The Tech Race
The senators' letter asks the White House to match AI-enabled attacks with AI-enabled defenses, placing federal cyber policy in direct competition with the same tools attackers now use. The disclosures from Anthropic and Microsoft show that competition is already running in active intrusions attributed to Chinese state-linked groups.
Organizations that depend on federal cybersecurity grants or staffing, including state and local governments, may see the effects of the staffing and grant reductions first. No availability dates or deployment timelines for AI-enabled federal defenses were announced in the source.
The takeaway
Watch whether the White House responds to the Warner and Rounds letter with a specific AI-enabled defense program or funding commitment. Also track whether Anthropic, Microsoft or OpenAI publish further disclosures of AI-automated intrusions attributed to state-linked groups.
Further reading
For background on how federal agencies and private firms defend against intrusions like these, see the Cybersecurity section.
Source note: This article includes information reported by Elko Daily Free Press.
Live Poll
Do you think the federal government is doing enough to protect against cyberattacks?





