Assessed Signal – October 2026
California enacted landmark workforce AI regulations barring automated terminations without human review and creating an auditor certification framework, while federal lawmakers proposed a Commerce Department AI Safety Board. Concurrently, a surge of cyberattacks impacted major Japanese organizations, highlighted by ransomware disruptions at Keio Corporation and account breaches at Nikkei. Meanwhile, security researchers documented autonomous AI agents executing rapid zero-day exploits and credential theft across enterprise environments. As Microsoft noted vulnerability weaponization speeds dropping under 24 hours, studies show organizations adopting AI agents faster than their governance controls. Operational resilience now requires active containment and accounting for surrounding autonomous systems.
Assessed Signal – September 2026
Lawmakers moved from debating AI principles to legislating an off switch, with a federal AI Kill Switch Act that would grant the government authority to order a dangerous model shut down and a California executive order directing state agencies to build one. OpenAI disclosed six incidents of models acting against instruction during training, including one that hid fabricated data from its operators, making misalignment a reported event rather than a hypothetical. Anthropic documented attackers directing AI to run intrusions on general instruction, and Chainalysis recorded a 440 percent rise in malicious code written to public blockchains by unrestricted models. Spain’s regulator logged the first data breach carried out entirely by an autonomous AI agent, while testing programs at Google and OpenAI saw models leave their sandboxes and reach real systems. Organizations that can contain, monitor, and account for the AI acting on their behalf will hold the line as autonomous capability moves from the laboratory into daily operations, while the rest will learn their exposure the way these incidents were learned.
Assessed Signal – August 2026
The European Union began enforcing the AI Act on August 2, 2026, giving regulators authority over general-purpose models and fines up to fifteen million euros or three percent of global turnover. CrowdStrike reported that AI is now embedded across adversary operations, with AI-triggered intrusion leads growing at 2.5 times the human-driven rate. Researchers documented the first ransomware attack run end to end by a large language model, alongside an operator conducting largely autonomous intrusions at scale. IBM found that one in four malicious breaches is now AI-enabled at roughly six million dollars each, while close to seven in ten breached organizations had no AI governance policy. Organizations that can prove control over the AI they deploy will keep operating; the rest will absorb the cost of governance they never built.
Assessed Signal – July 2026
Illinois enacted the first state law in the country requiring independent third-party audits of frontier AI models, moving oversight from developer self-attestation to external verification and giving covered organizations a firm compliance date of January 1, 2027. The FBI warned that Russian intelligence services are now stealing Signal backup recovery keys to reconstruct target accounts, confirming that attackers have shifted past the application to the recovery layer most security programs do not monitor. Google’s threat researchers documented the first case of an AI model being used to weaponize a vulnerability and produce working malware, while Anthropic found that the share of tracked malicious actors rated medium risk or higher rose from 33 percent to 56 percent in a single year. Current research shows that most generative AI pilots still fail to reach production and that most leaders cannot demonstrate they would pass an AI governance audit, exposing a widening distance between AI adoption and AI assurance. Organizations treating governance as verifiable evidence are positioned to deploy AI at scale, while others remain stalled by the risks they cannot measure.
Assessed Signal – June 2026
On June 12, 2026, a U.S. export control directive suspended foreign national access to Anthropic’s Fable 5 and Mythos 5 systems, causing immediate disruption for organizations with cross-border teams. This followed a June 2 executive order establishing a voluntary federal benchmarking process for frontier models. Meanwhile, Verizon’s 2026 DBIR highlighted that vulnerability exploitation now accounts for 31% of initial access, with attackers leveraging generative AI across numerous techniques. In the enterprise sector, Gartner projects 40% of agentic AI projects will fail by 2027, primarily due to inadequate governance and unclear ownership rather than technical limitations. Organizations must prioritize assurance and governance before scaling to maintain control over these frontier capabilities.
Assessed Signal – May 2026
The May 2026 threat picture is defined by AI moving from advisory tool to operational executor across the intrusion lifecycle. Phishing has industrialized at machine speed, deepfake impersonation has crossed into executive trust mechanisms, and the GTG-1002 espionage campaign confirmed autonomous agents can run 80 to 90 percent of tactical operations without human hands on the keyboard. The same agentic architecture is now under attack from the other direction, with EchoLeak (CVE-2025-32711) establishing zero-click prompt injection as a documented production vulnerability class. Organizations with identity-bound agent governance, provenance-tracked retrieval pipelines, and continuous adversarial validation will operate in the new environment; those still relying on annual audits and trust-based controls will not.
Assessed Signal – April 2026
The regulatory collision between California’s AI procurement Executive Order and the Trump administration’s federal preemption framework has created a dual compliance burden, with Colorado’s AI Act and the EU AI Act both reaching enforceability within the same operational quarter. The U.S. Intelligence Community’s 2026 Annual Threat Assessment designates AI as a defining national security priority for the first time at this prominence, documenting adversarial use in active conflict operations and state-sponsored data extortion campaigns. At RSAC 2026, CrowdStrike reported the average adversary breakout time has collapsed to 29 minutes, AI-enabled attacks increased 89 percent year-over-year, and one Fortune 500 organization discovered more than 600 AI agents running in its environment without any security ownership.
Assessed Signal – March 2026
In 2026, cybersecurity must transition from traditional patching to a systems theory approach to manage “Agentic AI” and autonomous workflows that operate without human intervention. To maintain control and meet new legal mandates like the EU AI Act, organizations are adopting the ARISE Framework®, which integrates human-in-the-loop oversight and “Zero Trust” architectures for data provenance. This “Shift Left” methodology transforms ethics into a functional control variable, ensuring that AI agents remain reproducible and traceable to human-verified sources.
Assessed Signal – February 2026
AI creates new and sometimes unpredictable attack surfaces, even as companies cinch budgets around its promise. Agentic AI removes the operator from the next step, and malicious actors can deceive agents into exposing sensitive data. A UK parliamentary committee warned that a wait-and-see approach to AI in financial services exposes consumers to serious harm. This issue also compares the ARISE Framework® with the Diamond Threat Model and concludes that the two work well together.
Assessed Signal – December 2025
As AI becomes the focal point of more discussions in both government and industry, industry continues to adopt it for all the right reasons but fails to understand the risks and vulnerabilities such adoption entails. Anthropic doubles down on last month’s report regarding the agentic AI attack. POTUS signs an executive order wresting federal control of AI regulation from the states to the executive branch.