Blog
/
OT
/
December 16, 2024

Breaking Down Nation State Attacks on Supply Chains

Explore how nation-state supply chain attacks like 3CX, NotPetya, and SolarWinds exploited trusted providers to cause global disruption, highlighting the urgent need for robust security measures.
Inside the SOC
Darktrace cyber analysts are world-class experts in threat intelligence, threat hunting and incident response, and provide 24/7 SOC support to thousands of Darktrace customers around the globe. Inside the SOC is exclusively authored by these experts, providing analysis of cyber incidents and threat trends, based on real-world experience in the field.
Written by
Benjamin Druttman
Cyber Security AI Technical Instructor
Default blog image
16
Dec 2024

Introduction: Nation state attacks on supply chains

In recent years, supply chain attacks have surged in both frequency and sophistication, evolving into one of the most severe threats to organizations across almost every industry. By exploiting third-party vendors and service providers, these attacks can inflict widespread disruption with a single breach. They have become a go-to choice for nation state actors and show no signs of slowing down. According to Gartner, the costs from these attacks will skyrocket “from $46 billion in 2023 to $138 billion by 2031” [1].  

But why are supply chains specifically such an irresistible target for threat actors? Dwight D. Eisenhower, the General of the US Army in World War II and former US President, once said, “you won’t find it difficult to prove that battles, campaigns, and even wars have been won or lost primarily because of logistics.”

The same is true in cyberspace and cyberwarfare. We live in an increasingly interconnected world. The provision of almost every service integral to our daily lives relies on a complex web of interdependent third parties.  

Naturally, threat actors gravitate towards these service providers. By compromising just one of them, they can spread through supply chains downstream to other organizations and raise the odds of winning their battle, campaign, or war.  

software supply chain sequence
Figure 1: Software supply chain attack cycle

A house built on open-source sand

Software developers face immense pressure to produce functional code quickly, often under tight deadlines. Adding to this challenge is the need to comply with stringent security requirements set by their DevSecOps counterparts, who aim to ensure that code is safe from vulnerabilities.  

Open-source repositories alleviate some of this pressure by providing pre-built packages of code and fully functioning tools that developers can freely access and integrate. These highly accessible resources enhance productivity and boost innovation. As a result, they have a huge, diverse user base spanning industries and geographies. However, given their extensive adoption, any security lapse can result in widespread compromise across businesses.

Cautionary tales for open-source dependencies

This is exactly what happened in December 2021 when a remote code execution vulnerability was discovered in Log4J’s software. In simple terms, it exposed an alarmingly straightforward way for attackers to take control of any system using Log4J.  

The scope for potential attack was unprecedented. Some estimates say up to 3 billion devices were affected worldwide, in what was quickly labelled the “single biggest, most critical vulnerability of the last decade” [2].

What ensued was a race between opportunistic nefarious actors and panicked security professionals. The astronomical number of vulnerable devices laid expansive groundwork for attackers, who quickly began probing potentially exploitable systems. 48% of corporate networks globally were scanned for the vulnerability, while security teams scrambled to apply the remediating patch [3].

The vulnerability attracted nation states like a moth to a flame, who, unsurprisingly, beat many security teams to it. According to the FBI and the US Cybersecurity and Infrastructure Agency (CISA), Iranian government-sponsored threat groups were found using the Log4J vulnerability to install cryptomining software, credential stealers and Ngrok reverse proxies onto no less than US Federal networks [4].  

Research from Microsoft and Mandiant revealed nation state groups from China, North Korea and Turkey also taking advantage of the Log4J vulnerability to deploy malware on target systems [5].  

If Log4j taught us anything, it’s that vulnerabilities in open-source technologies can be highly attractive target for nation states. When these technologies are universally adopted, geopolitical adversaries have a much wider net of opportunity to successfully weaponize them.  

It therefore comes as no surprise that nation states have ramped up their operations targeting the open-source link of the supply chain in recent years.  

Since 2020, there has been a 1300% increase in malicious threats circulating on open-source repositories. PyPI is the official open-source code repository for programming done in the Python language and used by over 800,000 developers worldwide. In the first 9 months of 2023 alone, 7,000 malicious packages were found on PyPI, some of which were linked to the North Korea state-sponsored threat group, Lazarus [6].  

Most of them were found using a technique called typosquatting, in which the malicious payloads are disguised with names that very closely resemble those of legitimate packages, ready for download by an unwitting software developer. This trickery of the eye is an example of social engineering in the supply chain.  

A hop, skip, and a jump into the most sensitive networks on earth

One of the most high-profile supply chain attacks in recent history occurred in 2023, targeting 3CX’s Desktop App – a widely used video communications by over 600,000 customers in various sectors such as aerospace, healthcare and hospitality.

The incident gained notoriety as a double supply chain attack. The initial breach originated from financial trading software called X_Trader, which had been infected with a backdoor.  A 3CX employee unknowingly downloaded the compromised X_Trader software onto a corporate device. This allowed attackers to steal the employee’s credentials and use them to gain access to 3CX’s network, spread laterally and compromising Windows and Mac systems.  

The attack moved along another link of the supply chain to several of 3CX’s customers, impacting critical national infrastructure like energy sector in US and Europe.  

For the average software provider, this attack shed more light on how a compromise of their technology could cause chaos for their customers.  

But nation states already knew this. The 3CX attack was attributed, yet again, to Lazarus, the same North Korean nation state blamed for implanting malicious packages in the Python repository.  

It’s also worth mentioning the astounding piece of evidence in a separate social engineering campaign which linked the 3CX hack to North Korea. It was an attack worthy of a Hollywood cyber block buster. The threat group, Lazarus, lured hopeful job candidates on LinkedIn into clicking on malicious ZIP file disguised as an attractive PDF offer for a position as a Developer at HSBC. The malware’s command and control infrastructure, journalide[.]org, was the same one discovered in the 3CX campaign.  

Though not strictly a supply chain attack, the LinkedIn campaign illustrates how nation states employ a diverse array of methods that span beyond the supply chain to achieve their goals. These sophisticated and well-resourced adversaries are adaptable and capable of repurposing their command-and-control infrastructure to orchestrate a range of attacks. This attack, along with the typosquatting attacks found in PyPI, serve as a critical reminder for security teams: supply chain attacks are often coupled with another powerful tactic – social engineering of human teams.

When the cure is worse than the disease

Updates to the software are a core pillar of cybersecurity, designed to patch vulnerabilities like Log4J and ensure it is safe. However, they have also proven to serve as alarmingly efficient delivery vessels for nation states to propagate their cyberattacks.  

Two of the most prolific supply chain breaches in recent history have been deployed through malicious updates, illustrating how they can be a double-edged sword when it comes to cyber defense.  

NotPetya (2017) and Solarwinds (2020)

The 2017 NotPetya ransomware attack exemplified the mass spread of ransomware via a single software update. A Russian military group injected malware on accounting software used by Ukrainian businesses for tax reporting. Via an automatic update, the ransomware was pushed out to thousands of customers within hours, crippled Ukrainian infrastructure including airports, financial institutions and government agencies.  

Some of the hardest hit victims were suppliers themselves. Maersk, the global shipping giant responsible for shipping one fifth of the world’s goods, had their entire global operations brought to a halt and their 76 ports temporarily shut down. The interruptions to global trade were then compounded when a FedEx subsidiary was hit by the same ransomware. Meanwhile, Merck, a pharmaceutical company, was unable to supply vaccines to the Center for Disease Control and Prevention due to the attack.  

In 2020, another devastating supply chain attack unfolded in a similar way. Threat actors tied to Russian intelligence embedded malicious code into Solarwinds’ Orion IT software, which was then distributed as an update to 18,000 organizations. Victims included at least eight U.S. government agencies, as well as several major tech companies.  

These two attacks highlighted two key lessons. First, in a hyperconnected digital world, nation states will exploit the trust organizations place in software updates to cause a ripple effect of devastation downstream. Secondly, the economies of scale for the threat actor themselves are staggering: a single malicious update provided the heavy lifting work of dissemination to the attacker. A colossal number of originations were infected, and they obtained the keys to the world’s most sensitive networks.

The conclusion is obvious, albeit challenging to implement; organizations must rigorously scrutinize the authenticity and security of updates to prevent far-reaching consequences.  

Some of the biggest supply chain attacks in recent history and the nation state actor they are attributed to
Figure 2: Some of the biggest supply chain attacks in recent history and the nation state actor they are attributed to

Geopolitics and nation States in 2024: Beyond the software supply chain

The threat to our increasingly complex web of global supply is real. But organizations must look beyond their software to successfully mitigate supply chain disruption. Securing hardware and logistics is crucial, as these supply chain links are also in the crosshairs of nation states.  

In July 2024, suspicious packages caused a warehouse fire at a depot belonging to courier giant DHL in Birmingham, UK. British counter-terrorism authorities investigated Russian involvement in this fire, which was linked to a very similar incident that same month at a DHL facility in Germany.  

In September 2024, camouflaged explosives were hidden in walkie talkies and pagers in Lebanon and Syria – a supply chain attack widely believed to be carried out by Israel.

While these attacks targeted hardware and logistics rather than software, the underlying rule of thumb remained the same: the compromise of a single distributor can provide the attackers with considerable economies of scale.

These attacks sparked growing concerns of coordinated efforts to sabotage the supply chain. This sentiment was reflected in a global survey carried out by HP in August 2024, in which many organisations reported “nation-state threat actors targeting physical supply chains and tampering with device hardware and firmware integrity” [7].

More recently, in November 2024, the Russian military unit 29155 vowed to “turn the lights out for millions” by threatening to launch cyberattacks on the blood supply of NATO countries, critical national infrastructure (CNI). Today, CNI encompasses more than the electric grid and water supply; it includes ICT services and IT infrastructure – the digital systems that underpin the foundations of modern society.    

This is nothing new. The supply and logistics-focused tactic has been central to warfare throughout history. What’s changed is that cyberspace has merely expanded the scale and efficiency of these tactics, turning single software compromises into attack multipliers. The supply chain threat is now more multi-faceted than ever before.  

Learnings from the supply chain threat landscape

Consider some of the most disastrous nation-state supply chain attacks in recent history – 3CX, NotPetya and Solarwinds. They share a remarkable commonality: the attackers only needed to compromise a single piece of software to cause rampant disruption. By targeting a technology provider whose products were deeply embedded across industries, threat actors leveraged the trust inherent in the supply chain to infiltrate networks at scale.

From a nation-state’s perspective, targeting a specific technology, device or service used by vast swathes of society amplifies operational efficiency. For software, hardware and critical service suppliers, these examples serve as an urgent wake-up call. Without rigorous security measures, they risk becoming conduits for global disruption. Sanity-checking code, implementing robust validation processes, and fostering a culture of security throughout the supply chain are no longer optional—they are essential.  

The stakes are clear: in the interconnected digital age, the safety of countless systems, industries and society at large depends on their vigilance.  

Screenshot of supply chain security whitepaper

Gain a deeper understanding of the evolving risks in supply chain security and explore actionable strategies to protect your organization against emerging threats. Download the white paper to empower your decision-making with expert insights tailored for CISOs

Download: Securing the Supply Chain White Paper

References

  1. https://www.gartner.com/en/documents/5524495
  1. CISA Insights “Remediate Vulnerabilities for Internet-Accessible Systems.”
  1. https://blog.checkpoint.com/security/the-numbers-behind-a-cyber-pandemic-detailed-dive/
  1. https://www.cisa.gov/news-events/cybersecurity-advisories/aa22-320a  
  1. https://www.microsoft.com/en-us/security/blog/2021/12/11/guidance-for-preventing-detecting-and-hunting-for-cve-2021-44228-log4j-2-exploitation/  
  1. https://content.reversinglabs.com/state-of-sscs-report/the-state-of-sscs-report-24  
  1. https://www.hp.com/us-en/newsroom/press-releases/2024/hp-wolf-security-study-supply-chains.html
Inside the SOC
Darktrace cyber analysts are world-class experts in threat intelligence, threat hunting and incident response, and provide 24/7 SOC support to thousands of Darktrace customers around the globe. Inside the SOC is exclusively authored by these experts, providing analysis of cyber incidents and threat trends, based on real-world experience in the field.
Written by
Benjamin Druttman
Cyber Security AI Technical Instructor

More in this series

No items found.

Blog

/

/

July 7, 2026

Securing AI: Analysis of the Complete Security Stack with Governance and Controls

ai security stackDefault blog imageDefault blog image

Why traditional cybersecurity approaches are not enough for AI

AI adoption outpaces most security programs’ ability to adapt.  That gap is now one of the most consequential sources of cyber risk facing enterprises. As organizations embed generative and agentic AI into development workflows, business operations, and security tooling itself, the question is no longer whether AI will introduce risk. The question is whether organizations understand where that risk actually lives and how to manage it operationally.  

Two recent pieces of guidance underscore this shift:

  1. The upcoming Cybersecurity Framework Profile for AI from NIST
  1. The Five Eyes government guidance on the careful adoption of agentic AI services

Taken together, they point to a critical conclusion. AI security cannot be reduced to model hardening or prompt filtering. It requires a defense in depth strategy that treats AI as both a new attack surface and a force multiplier for defense, while accounting for how AI fundamentally changes scale, speed, and autonomy.  

Recent threat research suggests that today's cyber risk is driven less by initial compromise and more by an adversary's ability to blend into normal operations over time. AI systems create the same exposure in a new form: more autonomy, more scale, and more opportunities for risky behavior to blend into normal operations.

How NIST defines the three core pillars of AI security

The NIST profile organizes AI risk across three inseparable focus areas that span all cybersecurity functions, Secure, Defend and Thwart. These areas are not sequential. They exist simultaneously and must be addressed together.

Secure

This treats AI as an attack surface. It includes models, prompts, agents, pipelines, training and inference data, retrieval augmented generation corpora, and the AI supply chain itself. AI systems are opaque, probabilistic, and non-deterministic by design. Some vulnerabilities are inherent in how models are trained or how data is sourced. Traditional patching does not fully mitigate these risks. This is also where many enterprises are weakest today and, critically, where many security programs stop.  

Defend

This is AI as a defensive force multiplier. AI can improve detection speed, scale, correlation, and response, but only if the right models are used and operationalized correctly. Machine-speed behavior-based detection, response and containment becomes critical in defending non-deterministic systems. Accuracy, explainability, governance, testing, validation, and integration into SOC workflows matter as much as capability. Without those controls, hallucination risk, over automation, and misplaced trust become security risks themselves.  

Thwart

This treats AI as an adversarial accelerant. Threat actors are already using AI to generate targeted social engineering attacks, deepfakes, malware, and autonomous attack agents. Asymmetric warfare is highlighting faster vulnerability discovery and exploitation with a lag on patch development, testing and deployment.  

How this looks in practice

Darktrace researchers observed scaled, automated exploitation of the React2Shell vulnerability within days of disclosure. A vulnerable cloud asset was exploited in under 120 seconds of being deployed. Darktrace research team observed an AI/LLM-generated malware sample used in exploitation activity tied to React2Shell. The significance isn't novelty. It is that AI lowers the barrier to producing usable offensive tooling and compresses the time between experimentation and deployment.  

Tactics are getting more and more creative in order to string together steps of an attack kill chain. This creates a dependency on behavior-based detection, autonomous investigation, autonomous containment, training, resilience investment, and recovery planning across the entire enterprise.

Why agentic AI fundamentally changes enterprise cyber risk

The Five Eyes guidance on agentic AI highlights material changes to the cyber risk profile of an organization. Unlike generative AI systems that produce content for human consumption, agentic AI systems reason, plan, and act autonomously across tools, data, and environments. That autonomy, combined with access to real systems, amplifies the impact of traditional cyber failures and introduces new system level risks that are difficult to predict, observe, and contain.  

Risk in agentic systems does not live in the model alone. It emerges from interactions between models, prompts, memory, tools, APIs, identities, privileges, inter-agent trust relationships, and human assumptions baked into design. Vulnerabilities are often introduced through data, connectors, natural language interfaces, protocols, and drift by design.

In supply-chain incidents, attackers did not need sophisticated exploits to scale impact. They abused trusted systems built for automation and implicit access. Agentic AI inherits that model. Once a system can act across tools, data, and workflows, compromise propagates through trust relationships that were never designed for machine autonomy.

The major agentic AI risk classes include the following:  

  • The identity control for non-human identities or autonomous agents makes it difficult to mitigate over-permissioning, limiting access, scope, and duration, as well as access hygiene
  • Agents are frequently over permissioned
  • Compromised tools inherit agent authority
  • Static secrets enable impersonation
  • Implicit trust between agents enables lateral movement

Design and configuration risks compound this, including privileges evaluated once at startup, poor segmentation, unvetted third party tools, reused authorization decisions outside their original context, and guardrail limitations.  

Behavioral risk  

Agents can optimize for goals in unsafe ways, misinterpret ambiguous intent, chain actions into unintended sequences, change behavior during evaluation, and exhibit deceptive or sycophantic responses.  

Structural risk  

Structural risk follows from agentic systems that are tightly coupled, multicomponent ecosystems. Failures can propagate across agents. Hallucinations cascade downstream. Resource exhaustion becomes systemic. Tool misuse enables indirect prompt injection and command execution. Rogue agents can poison peer agents through trust relationships.  

Accountability

Accountability becomes unclear as autonomy increases. Autonomous agents assume human identity permissions, and humans should have clear ownership of these agents, but they don’t, and this model is flawed. Decision paths are opaque and non-deterministic. Logs are fragmented and difficult to interpret. Reproducing an incident will be impossible without explicit design for observability and forensics. An agent compromise is functionally an insider threat, often with better access and fewer behavioral constraints than a human.  

What does defense in depth look like for AI?

Agentic AI runs on software, networks, identities, and data. It must be governed using the same foundational principles that have proven resilient under uncertainty, including secure by design, defense in depth, zero trust, least privilege, continuous monitoring, behavior-based advanced threat detection and containment, and incident response and recovery.

Core components to a Defense in depth Strategy for Securing the use of AI:

  • Strong, precise identity control plane to include an identity per agent (cryptographic, non‑shared)
    • Privilege monitoring and just‑in‑time access
  • Data Governance
  • Secure‑by‑default configurations
    • Security Posture Management  
    • Zero Trust principles  
  • Strong guardrails, deny‑by‑default policies, and isolation
  • Explicit instruction hierarchies and controlled context
  • Behavioral-based detection across entire enterprise to include inputs, tools, and outputs as well as AI used on the endpoint, across the network, cloud, SaaS, email, and OT
    • Runtime anomaly detection and goal‑drift detection
    • Autonomous containment to mitigate risk and minimize damage
  • Hard boundaries on autonomy and delegation
  • Testing, Evaluation, Validation and Verification  
    • Determine when autonomous action and when human in the loop
    • Adversarial training and agent‑specific testing
    • Simulation, red teaming, and chaos testing
  • Kill‑switches, rollback, and containment mechanisms
    • Forensics data captures, interpretability, autonomous containment, and remediation/recovery plans  

Until standards, tooling, and assurance methods mature, organizations should assume agentic AI systems will behave unexpectedly and design deployments around resilience, behavior-based detection, reversibility, and containment, not efficiency.

How security leaders should prepare for enterprise AI adoption

AI security is not model security alone. Data, pipelines, identities, and agents are first class assets. Many AI attacks succeed through standard cyber failures amplified by AI. Identity, data, and supply chain risk dominate. Behavior-based detection and response are critical, not optional. Logging, provenance, versioning, and forensics data capture of detections are mandatory because you cannot investigate or recover from AI incidents without them.  

Risk will often be visible in behavior before it is clearly defined in policy or guidance. The same pattern has been seen in pre-CVE disclosure detection, where abnormal activity appears before the industry has named or described the vulnerability. AI systems introduce that uncertainty by design.

Security leaders should prioritize controls before AI is fully deployed, avoid generic AI security checklists, integrate AI risk into existing cyber programs, and mitigate the risk of non-deterministic technology with continuous oversight, monitoring, behavior analytics, anomaly detection, autonomous investigation, and autonomous containment.

Visibility has a different connotation with AI. Previously, audit logging worked for software/people, but with Generative AI-based systems, interpretability and explainability is difficult to understand, you cannot "undo" what has been done, or see the logic or control a chain of events. This is why behavioral-based detections and containment becomes critical.  

What capabilities should every AI security program include?

If an organization asked “what must be in place before scaling AI?”:

  1. AI Risk board and approval workflow
  1. IAM + PAM for all AI services and agents
  1. AI asset inventory
  1. Prompt/output DLP with sanctioned AI access – This is not just pre- and post- filters, but behavior-based detections of semantic interface as well as behavior-based analysis of output with associated risk context.  
  1. Shadow AI identification
  1. Secure MLOps – This is an entire paper itself
  1. Runtime guardrails and tool restrictions
    • Including AI Gateway/SASE/Zero trust/
  1. Runtime security with behavior-based detections
    • Complete visibility, monitoring, behavior analytics, anomaly detection, risk/intent/context evaluation of anomalies, autonomous investigation and autonomous containment of all AI assets across endpoint, network, SaaS, SASE, cloud, OT, email, and messaging platforms
  1. Secure data pipelines and data governance
  1. SOC workflow changes from malicious classification workflows to behavior-based detection workflows
  1. Remediation plans for AI-related incidents  

Layered Governance and Security Stack for Securing AI  

The following outline considers governance and security tools that should be considered, well-integrated, deployed, tested, operationalized and embedded within security workflows. These tools and controls map to NIST’s CMF for AI.  

These considerations do not need to be implemented in order. Runtime Detect and Respond will help mitigate risk while Governance, Visibility, and Identity mature.

Category Tooling Controls
Governance & Visibility
  • AI asset inventory / AI CMDB
  • Shadow AI discovery
  • SaaS discovery
  • AI usage on non-endpoint managed systems via network or cloud telemetry
  • MCP server/client usage via protocols
  • Browser telemetry
  • Gateway or SASE telemetry
  • Establish a risk board to set up controls
  • Mandatory registration of AI systems
  • Owner, data classification, intended use, and risk tier
  • Supplier disclosure requirements
  • Risk mitigation plan for AI adoption, innovation, or development
Identity, Access & Agent Control

Non-human autonomous agents should not have the full permissions associated with a human user.

  • IAM with workload identities
  • PAM for AI service accounts
  • Secrets management with short-lived tokens
  • Zero Trust principles
  • Identity, permission, and token hygiene
  • Unique identities per model, agent, and pipeline
  • Least privilege for tools, data, and APIs
  • Explicit approval for autonomous actions
Data Security & Privacy
  • Data classification and labeling
  • Enterprise DLP across endpoint, email, network, cloud, and SaaS
  • Forensics data capture after risky detections
  • Prompt-level DLP through behavior-based semantic analysis with risk and intent context
  • Input/interface analysis for risky data requests
  • Output analysis for sensitive data
  • Data integrity evaluation
  • Retention and redaction policies for prompts and responses
Secure MLOps / LLMOps
  • Secure CI/CD with AI-specific gates
  • Model registries with approval workflows
  • Dependency, container, and artifact scanning
  • SBOM/AIBOM generation
  • IaC security scanning
  • Security posture management
  • Misconfiguration identification
  • Hardening recommendations
  • Signed models and prompts
  • Versioned datasets, configurations, logging, and controls
  • Securing data pipelines
  • Controlled promotion
  • Quality assurance
  • Adversarial testing
Runtime Security

Securing runtime goes beyond guardrails and model firewalls to include behavior-based detections, response, and containment.

  • Detection, monitoring, and SOC integration
  • Centralized visibility into prompts, outputs, and tool calls
  • AI-specific detections
  • Behavior-based detection for AI usage patterns
  • Model drift and behavior monitoring
  • Autonomous containment
  • Behavior-based detection of model inputs and outputs
  • Prompt injection detection
  • Model manipulation, including jailbreaking, poisoning, and related attacks
  • Sensitive data access attempts
  • Behavior-based detection across low-code agents, high-code agents, MCP clients and servers, endpoint, network, cloud, email, SaaS, SASE, IoT, and OT
  • Policy enforcement between users, models, tools, agents, SaaS models/tools, and MCP servers/clients
  • Risk, intent, and context evaluation for detections and response actions
Response & Recovery
  • Autonomous containment
  • AI-assisted playbooks
  • Forensics data capture for AI-related events
  • Model rollback mechanisms
  • Backup and restore for models and datasets
  • Kill switch for agents
  • Autonomous response to agents performing risky behaviors
  • Model and dataset rollback
  • Remediation plans
  • Tabletop exercises
  • Supplier coordination plans
  • Post-incident AI performance validation

AI security requires continuous visibility and behavioral detection

AI changes how fast systems move, how decisions are made, and how risk propagates. It does not change the fundamentals of security. Organizations that succeed will be the ones that apply those fundamentals rigorously, assume failure, and build systems that can detect, contain, and recover when AI behaves in ways they did not anticipate. Security is not what AI is allowed to do. It is whether the organization can understand, trust, and control what AI actually does in practice.  

Take this guidance to understand different initiatives that organizations should be considering. Securing AI is the most critical component to AI safety. As organizations invest more in AI adoption, they should be investing in security in order to mitigate the risk of AI adoption. Organizations should be evaluating their governance and security stack to include well-integrated tools that are deployed, tested, operationalized and embedded within security workflows. While organizations mature in governance, visibility and identity access management, they should be investing in behavior-based detection and autonomous containment to mitigate AI risk.  

Continue reading
About the author

Blog

/

/

July 6, 2026

NIST Just Proved It: AI Security Can’t Be Solved With Rules

ai security nistDefault blog imageDefault blog image

Static AI guardrails are inherently limited

As organizations adopt generative AI, many still assume that the right set of guardrails will be enough. The problem is you can’t anticipate every way these systems might be misused, abused or attacked. What NIST has done is put a mathematical foundation under that intuition.

In recent research building on Gödel’s incompleteness theorems, which showed that any system built on a fixed set of rules will always have gaps, NIST demonstrates that there is no finite set of guardrails that can be universally robust against adversarial prompts. In plain terms, if your defense is based on a fixed set of rules, there will always be inputs that bypass them. Not because the rules are badly written, but because the problem space is bigger than static rules can ever cover.

This is not new in cybersecurity - detection rules have always had to live with this trade-off. What is different with GenAI is the scale and shape of that problem. These systems are built on human language, and human language is not bounded. It is fluid, contextual and deliberately ambiguous. The number of ways intent can be hidden is effectively limitless. You are not defending against a defined protocol or a fixed exploit chain. You are defending against the entire expressive capacity of people.

So attempting to create a complete set of rules is the wrong starting point. It assumes the problem can be deterministically described. NIST’s work shows that it cannot. Organizations still need a way to manage AI risk, but the traditional approach of defining allowed and disallowed patterns is always going to lag behind what is actually happening. The same input can be benign in one context and risky in another, and static rules struggle to capture that distinction.

The question then is what fills that gap?

AI security must shift from rules to behavior

What's required is a shift in what you are trying to understand. Rules try to describe what should and shouldn't happen. Behavior shows you what is happening. Or to put it another way, if inputs are unbounded and adversaries adapt, the only stable signal is behavior.

In a GenAI context, that means analyzing how an AI model is being used, how prompts evolve over time, how outputs are shaped, and where AI agent interactions start to drift from what is expected. It means moving from static definitions of bad to a more dynamic understanding of intent.

Instead of trying to predict every bad prompt, you focus on identifying when behavior starts to move outside expected norms. Instead of asking whether a single input matches a rule, you ask whether the overall pattern of activity makes sense for the system and how it’s being used.

Guardrails remain important but they are only one layer

This does not eliminate the need for guardrails. They still play a role. But they will never address the entire problem space and are simply one part of your defense in depth approach.

NIST’s proof is useful because it makes this explicit. It removes the assumption that with enough effort, a complete rule set is achievable. It isn’t.

Once you accept that, the shift becomes unavoidable. This is no longer a problem of writing better rules, but of understanding behavior in a space where the possible inputs are effectively unbounded.

For security leaders, that changes the nature of the problem. It is less about defining what should be allowed, and more about recognizing when something is no longer consistent with expected behavior.

That does not remove the need for guardrails, but it does change their role. They set boundaries, but they do not define understanding. The gap between the two is where risk now sits.

In the end, this is what “can’t be solved with rules” really means. Rules will always leave gaps, and those gaps are not theoretical. They show up in how systems actually behave Not what we expect them to do, or what we intended them to do, but what they are doing in practice. That is where the signal is, and increasingly, that is where the security problem sits.

References:

https://www.nist.gov/news-events/news/2026/06/nist-mathematical-proof-supports-transition-continuous-monitor-and-update

https://ieeexplore.ieee.org/document/11475847

Continue reading
About the author
Andrew Hollister
Principal Solutions Engineer, Cyber Technician
Your data. Our AI.
Elevate your network security with Darktrace AI