ブログ
/
OT
/
June 11, 2026

Protecting Stadiums & Events with AI

A stadium during a major sporting event is simultaneously a city, a broadcast hub, a transport node, a public safety space, and a national symbol. That makes them attractive not just to criminals, but to politically motivated and state‑aligned threat actors willing to plan years ahead. With the 2026 FIFA World Cup spanning three nations and dozens of host cities, those challenges – and the AI now amplifying them – have never carried higher stakes.
No items found.
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.
No items found.
Default blog image
11
Jun 2026

Stadium and large public venue operators are confronted with a unique set of cyber security challenges. Often described as a ‘honeypot’ for cyber-criminals, the sports and entertainment industry is an attractive target for threat actors for three main reasons:

  • Modern sports organizations process sensitive and highly valuable data at scale;
  • Sporting events are highly visible and time-critical, operating in front of live audiences with no room for error;
  • Sports organizations rely on sprawling vendor ecosystems and supply chains to deliver broadcast, commerce, fan engagement services, and more.

In a recent Darktrace-commissioned survey, 84% of professional sports organizations reported at least one cyber incident in the past year, and 57% were hit more than once [1]. The potential ramifications of cyber disruption during a large-scale sports event cannot be overstated. A momentary lapse in access to power could bring TV broadcasts to a halt; disruption to access controls could restrict fans from entering the grounds; CCTV outages could increase the risk of criminal behavior and physical injuries. If data is not reliable and stadium machines are outputting the wrong metrics, a venue could become dangerously overcrowded. The barrier between the cyber and physical worlds has long dissolved – cyber-attacks threaten human safety.

In this blog, I explore the key challenges of stadium cyber security and explain the unique capabilities of Self-Learning AI that led me to adopt Darktrace as a head of ICT and cyber security for international venues and events. Over my career I have helped secure football and rugby World Cups, World Athletics Championships and more than 500 events ,and the lessons from each have only sharpened my conviction in this approach.

The access paradox

The biggest challenge lies in the paradox of securing a site where various internal services are provided to a large number of unknown and unmanaged users, suppliers and devices. When it’s game time, or ‘D-Day’, you see a huge influx of thousands of people, each with their own devices, needing to connect to your network and your infrastructure. The floodgates are opened. But certain parts of your digital environment need to remain protected: your sensitive employee and customer data, your critical OT systems. I liken this to opening the door to your home, and letting the entire town come in and wander around. But you still need to secure your master bedroom.

A multitude of different actors must be able to work on-site to provide services or content during the event. Broadcasters, staff and suppliers need to have access to manage the show, and all these people need to access or interact with the IT infrastructure. In many ways, these additional bodies are already inside the perimeter and could host unknown malicious threats.

This year, the paradox is wider than ever. A tournament spread across hundreds of suppliers and vendors means the foothold an attacker needs may already belong to a trusted partner – a single compromised supplier can become the doorway to everything else. And the adversary is no longer working alone: generative AI now lets attackers probe and weaponize vulnerabilities across thousands of software dependencies at a speed no human team could match, turning the access paradox from a manageable risk into a fast-moving target.

Achieving this balance between accessibility and security requires a shift in mindset from perimeter-based security to one that can detect and respond to threats on the inside. The complexities involved requires technology that can identify malicious behavior in real time based on the wider context of an incident. A particular behavior or connection may be benign in one context and yet critically disruptive in another — tools and technology must be able to discern between the two.

This is why I considered Darktrace’s Self-Learning AI a suitable fit: rather than defending at the perimeter, it focuses on detecting and responding to malicious activity already inside. Because it learns the unique ‘patterns of life’ of its surroundings, it can detect subtle deviations that indicate a threat and initiate a targeted response – without relying on pre-programmed rules and playbooks.

IT/OT convergence

The second key challenge is the issue of IT and OT convergence. Typical stadiums and arenas consist of a wide range of Industrial Control Systems (ICS).

This involves a complex and messy array of switches, cables, CCTV cameras, as well as devices and technologies being brought in by the media and the press, and all these IT and OT components are now interconnected, which means these technologies now have Internet Protocol (IP)-based threats to manage. The same challenges that the corporate infrastructure for stadium management faces in cyber security are therefore also now an issue for ICS security.

This challenge cannot be addressed by viewing IT and OT security in isolation — these two environments are linked because of the analogue migration to IP. A unified approach is required to detect and respond to threats that start in IT before moving to industrial systems.

The stakes are physical. CCTV, Access Control, Public Annoucement system, lighting and the giant screens are all now running over IP, and a disruption to any of them can force a venue to halt play on safety grounds. Scale compounds the problem. At the Qatar 2022 World Cup, eight stadiums were purpose-built to a single technical standard, which made the digital environment relatively uniform to defend. The 2026 tournament is the opposite: dozens of host venues across three countries, each with its own operator, its own contractors and its own legacy systems.This creates a far more fragmented and unpredictable estate to secure.

In addition, cyber security technology must be able to deal with complexity. Darktrace’s AI thrives in the most complex environments, with more data points adding more context to inform the AI’s decision making. It covers OT and IT with a single, unified AI engine, that can also detect and respond across cloud infrastructure, SaaS applications, email systems and endpoints. It is ready to adapt to the messy, interconnected systems that make up large stadiums’ digital infrastructure.

The time factor

Finally, the nature of stadium events means that timing is critical and puts enormous pressure on the organizers and operators. ‘D-Day’ cannot be replayed or postponed, and so if cyber disruption occurs during the event, every minute is crucial. You cannot reschedule a World Cup final or move an opening ceremony; the date is fixed, the world is watching, and there is no second take.

There is consequently a strong emphasis on two key metrics

  • Mean Time To Know (MTTK) — how long it takes the security team need to be aware of an incident; and
  • Mean Time To Restore (MTTR) — how quickly a team can act to contain the threat.

It is perhaps more imperative in stadium event management than anywhere else that these two metrics be minimized.

This leads to the third criteria in assessing cyber security technology: does it help with response? And critically, can that response be nuanced and targeted, able to contain that threat without causing further disruption?

To this end, Darktrace’s Autonomous Response takes machine-speed action to contain cyber-attacks, when humans are too slow to react or aren’t around at all. It’s powered by Darktrace’s AI, so it has a nuanced and continuously updating understanding of what’s ‘normal’ across IT and OT systems. This means its response actions are targeted: designed to eliminate the threat, but not at the cost of disruption. Crucially, this enables responses that are surgical rather than blunt. For example, taking an entire server offline to stop a ransomware attack can cause more disruption than the attack itself, so the real value lies in neutralizing the malicious activity precisely — containing the threat without taking down the systems the event and business depends on.

Depending on the nature and severity of the threat, the technology can block specific malicious connections by enforcing the normal ‘pattern of life’ of a device or account. When every second counts, this is the speed and granularity that you need in a cybersecurity technology.

Darktrace can be deployed across every area of the digital enterprise, including network, email, cloud and SaaS environments with the same self-learning approach, stopping anomalous behaviors that point to account takeover and other cloud-based threats. Earlier this year, we announced that Darktrace is also extending its behavioral approach to help businesses deploy and scale AI securely by understanding how these AI systems and agents behave, interact with other systems and humans, and evolve over time. This is critical because 72% of cybersecurity professionals at sports organizations believe AI will increase their cyber risk over the next 12 months [2].

Wherever it is deployed, Darktrace allows the stadium operator to focus on the vital part of the game and offers real-time protection without any modification in the network topology or infrastructure.

An adaptive defense

Cyber-criminals are constantly developing their approach in an attempt to evade security tools trained to look for specific hallmarks of an attack. As they get creative and continuously experiment with new tactics and techniques, the human operators using these tools are forced into a constant state of catch up.

An AI-based approach that learns an organization and its normal behavior patterns from the ground up puts an end to this game of ‘cat and mouse’, shifting the balance in favor of the defenders and allowing them to stay ahead of the threat. This matters more than ever, because adversaries are now using AI to scale their attacks. If you do not have AI working to protect you against malicious AI, you are already at a disadvantage.

With a nuanced understanding of what’s ‘normal’ for the business, unified IT/OT coverage, and an Autonomous Response solution that takes immediate, surgical action, the playing field is leveled, and large stadium and events operators can focus on delivering the best possible experience for attendees, digital viewers, partners and performers.

[related-resource]

References:

[1] [2] Darktrace: Cybersecurity in Global Sport, June 2026. Findings based on survey of 875 IT cybersecurity professionals based in the US, UK, Australia and Germany, working in professional sports organizations (including clubs, societies & sporting bodies) employing 10+ people. The survey was fielded between May 28, 2026 and June 3, 2026 by independent market research agency, Opinion Matters.

Cybersecurity in Global Sport

This report examines the threats, signals, and strategic implications for an increasingly digitized sports industry.

No items found.
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.
No items found.

More in this series

No items found.

Blog

/

AI

/

July 10, 2026

AIインフラがアタックサーフェスの一部に

Default blog imageDefault blog image

AIインフラとアタックサーフェスの進化

多くの組織が生成AIを実運用環境に導入するなかで、企業のクラウド環境内に新たなインフラのレイヤーが出現しています。それはAIゲートウェイです。AIゲートウェイはユーザー、アプリケーション、基盤モデルの間に位置し、多くの場合クラウドの特権アクセスを保持し、さまざまなAIサービスへのアクセスを大規模に管理しています。

AIゲートウェイとは?

AIゲートウェイはユーザー、アプリケーション、基盤モデルの間に位置し、多くの場合クラウドの特権アクセスを保持し、さまざまなAIサービスへのアクセスを大規模に管理しています。

こうした役割から、AIゲートウェイは企業のアタックサーフェスのますます重要な一部になりつつあります。AIゲートウェイが侵害されれば、攻撃者に対して計算リソースへのアクセスだけでなく、クラウドアイデンティティ、モデルサービス、機密性の高いプロンプト、そして他の接続されたシステムへのアクセスも提供してしまいます。

このブログでは、Amazon Bedrock サービスに接続されたAIゲートウェイが侵害され、その後暗号通貨マイニングインフラとの通信が観測された事例をダークトレースがどのように調査したかを解説します。問題のインスタンスは、その構成、ならびに関連するIAM(Identity and Access Management)ロールから、Amazon BedrockでホスティングされるAIサービスへのゲートウェイとして機能していることがわかりました。疑わしい侵害アクティビティが発生した後、このホストは既知の暗号通貨マイニングインフラに繰り返し通信を行い、その後シャットダウンされた様子が観測されました。Darktrace はこのアクティビティを検知し、Enhanced MonitoringおよびManaged Threat Detectionサービスを通じてエスカレーションを行いました。

この事例では最終的影響は不正な暗号通貨マイニングでしたが、このインシデントが注目に値するのはその発生場所です。侵害されたアセットは、クラウドインフラ、アイデンティティ、各種AIサービスの交差する場所に位置していました。最近の調査では、LiteLLM等のAIゲートウェイが、認証情報、モデルへのアクセス、クラウド権限を中央管理するその能力から、攻撃者にとって魅力的な標的となる可能性が明らかになっています。このアクティビティと公開されているLiteLLM脆弱性を直接結びつける証拠は見つかっていませんが、このインシデントは、AIインフラを個別のアプリケーション層として見るのではなく、重要なアタックサーフェスの一部として扱う必要性があることを表しています[1]。

暗号通貨マイニングがクラウド侵害後のアクティビティとしてよく見られる背景

暗号通貨マイニングはクラウド環境において、侵害後のアクティビティとして収益性の高いものとなり得ます。クラウド資産にアクセスできるようになった後、攻撃者はマイニングソフトウェアを展開して被害者の計算リソースを悪用し金銭的利益を得ることができます。この種のアクティビティは多くの場合機会主義的なものであり、露出したサービス、弱い認証情報、漏洩したアクセスキー、脆弱なアプリケーション、あるいはクラウドワークロードの設定ミスなどを標的として実行されます。

典型的なクラウド上での暗号通貨マイニング侵入には次のようなアクティビティが含まれます:

  • 露出したあるいは脆弱なクラウドインフラの特定
  • 露出したサービス、認証情報、またはアプリケーションの脆弱性を通じたアクセスの獲得
  • マイニングソフトウェアのダウンロードおよび実行
  • マイニングプールインフラへのアウトバウンド接続を繰り返し確立
  • アクティビティが検知され停止されるまで継続して計算リソースを消費

この事例において注目すべき要素は暗号通貨マイニングだけではありません。それが発生した場所が、AI関連アクティビティをサポートするクラウドインフラ上だったことです。この事例は、AIサービスを実現するためのアセットも、よくあるクラウド侵害リスクにさらされる可能性があることを示しています。

Amazon Bedrockに接続されたAIゲートウェイの侵害を調査

2026年6月12日、DarktraceはLiteLLM-Proxyという名前のAmazon Web Service (AWS) EC2インスタンスから暗号通貨マイニング発生中とみられるアクティビティを観測しました。このインスタンスはLiteLLMアクティビティをサポートしており、Amazon Bedrockリソースへのアクセス権を有するインスタンスプロファイルと関連付けられていました。  

AIゲートウェイは大規模言語モデルへのアクセスを中央管理するよう設計されており、多くの場合AIアプリケーションに対する認証、ルーティング、ログ、ポリシー適用を扱っています。セキュリティの視点から見ると、クラウド権限、モデルアクセス、アプリケーションワークフローを単一の制御ポイントに集約する役割も果たしています。その結果、AIゲートウェイの侵害は、侵害されたホストだけにとどまらない影響を及ぼす可能性があります。

確定的な初期アクセスベクトルは確認できませんでしたが、このアクティビティはインターネットに接続されているシステムの侵害でよく見られる次のような順序に従っていました。ブルートフォースアクセス、ペイロードの投下、そしてマイニングプールインフラに対する繰り返しのアウトバウンド接続です。

ステージ1: インターネットに露出したSSHからの初期アクセス

暗号通貨マイニングアクティビティが観測される前、LiteLLM-Proxy EC2インスタンスはSSH(ポート22)が0.0.0.0/0に対して開かれ、外部に公開されていました。

図1:EC2インスタンスがSSHポート22に対してすべてのインバウンドトラフィックを許可している設定ミスをDarktraceが警告

暗号通貨マイニングアクティビティに先立って、Darktraceはこのインスタンスに対する大量のインバウンド接続の試みが外部IPアドレス(主に145.241.123[.]102)からポート22に対して行われていることを観測しました。これはブルートフォースアクティビティを示唆するものです [2]。これらの接続の多くは短命であり、数秒しか続いておらず、スキャニングまたはログインの失敗を示していました。

図2:Darktraceがデバイスのポート22に対する不審なインバウンド接続試行を検知

入手できたテレメトリーではこれらのインバウンドSSH接続のいずれかが認証の成功につながったかどうかの確認に至らず、このアクティビティが初期アクセスベクトルであると断定することはできませんでした。しかしながら、SSHの露出、外部IPアドレスからのインバウンド接続、それに続くマイニングアクティビティは、SSHがアクセス経路の可能性が高いことを示唆しています。

ステージ2: AIゲートウェイへのXMRigマルウェアのダウンロード

最初に観測されたマイニングプールへの接続の後、このEC2インスタンスは3.42 MBのデータをポート80上のHTTP接続を介して外部エンドポイント185.62.1[.]8にダウンロードしました。このエンドポイントは暗号通貨マイニングマルウェアXMRigを含むZIPファイルをホスティングしていました[3][4]。ホストレベルのログは入手できなかったため、ダークトレースはマイニングツールがどのように実行されたか、あるいは前のSSHアクティビティがペイロード投下を直接的に可能にしたかどうかを確認できませんでした。しかしながら、ダウンロードのタイミングとその後ほどなくマイニングプールへの接続が繰り返されたことは、このインスタンスが侵害されて不正な計算アクティビティに使われたという評価を裏付けています。

ステージ3 – 侵害されたAIゲートウェイが暗号通貨マイニングインフラと通信

わずか数分後、DarktraceはLiteLLM-ProxyEC2インスタンスがHTTPs(ポート443)でホスト名pool.hasvault[.]proに対して接続していることを確認しました。最初の接続の後、同じホスト名に対して繰り返しアウトバウンド接続が観測されました。これは、侵害されたホストがマイニングインフラと通信しワークを受け取り、結果を送信するという、暗号通貨マイニングプールとの通信のパターンと一致しています。

このアクティビティがDarktraceのEnhanced Monitoringモデル“Compromise / HighPriority Crypto Currency Mining”をトリガーし、ダークトレースのSOCにより顧客に対してエスカレーションされました。また、このアクティビティはCyber AI Analystによって分析され、関連するイベントが1つの調査ナラティブにまとめられました。これにより、影響を受けたクラウドアセットからマニングプールへの繰り返しの接続を特定することができました。

図3:CyberAI Analystによる暗号通貨マイニングアクティビティの調査  

ポート443上のHTTPSの使用にも注目すべきです。なぜならば、単独で見れば、このトラフィックそのものは疑わしく見えないかもしれないからです。しかしこのケースでは、接続先、接続の量、そして類似のアクティビティが他にないことなどが、この通信を疑わしいものとして特定するのに必要な、動作のコンテキストを提供することになりました。

ステージ4: Managed Threat Detectionサービスによるリソース乱用の特定

暗号通貨マイニングアクティビティがダークトレースのManaged Threat Detectionサービスにより検知され、ダークトレースのSOCによりレビューされました。レビューの結果、このアクティビティは顧客向けにエスカレーションされました。このエスカレーションにより、顧客はAWS環境で現在発生中のリソースの乱用について、タイムリーな通知を受けることができました。

ステージ5: クラウド認証情報の不正使用とみられる疑わしいIAMアクティビティ

これとは別に、6月13日、Darktraceは別のIAMユーザーから発生した疑わしいアクティビティを検知しました。

図4: DarktraceのAdvanced Search機能が別のIAMユーザーが実行した疑わしいアクティビティをハイライト

まず、このユーザーは “GetSendQuota”イベントを試行している様子が見られました。このアクションは少なくとも過去3か月間にこのアカウントによって実行されたことのないアクションです。また、このコマンドのソースIPアドレスは14.176.1[.]47でした。地理位置情報はベトナムであり、このユーザーのアクティビティがAmazon IPアドレスから最も多く見られた場所です。さらに、このアクティビティに対してAWS CLIが使用されており、これもこのユーザーにとって通常とは異なる振る舞いでした。このことは、Darktraceの“IaaS / Unusual Activity / UnusualAWS CLI Activity”モデルによって検知されました。

図5: Darktraceによる “GetSendQuota” イベントの検知

このIAMユーザーからは、長期アクセスキーを使った疑わしいアクティビティがさらに観測されました。中でも、“InvokeModel” および “ListFoundationModels”コマンドの失敗が検知されており、モデル列挙や起動などAmazon Bedrockサービスとのやり取りを試行したことがわかります。これは前日観測されたLiteLLM侵害への関連を思わせますが、2つのイベントを確定的に結びつける証拠は不十分でした。

“CreateUser”コマンドの試行も注目に値します。なぜなら要求されたユーザー名は意味が薄いものであり、新しいアカウントを作成することにより永続性を確立する試みと見られるからです。このアクティビティはDarktraceのモデル“IaaS / Admin / New AWS UserAccount Creation”をトリガーしました。

図6:Darktraceによる“CreateUser” イベントの検知

2つのインシデント間に結びつきは確認できなかったものの、このIAMアクティビティには重要な意味があります。これは、クラウド侵害の調査においてワークロードのテレメトリーとコントロールプレーンのテレメトリーの両方を取り入れることの重要性を表しています。EC2暗号通貨マイニングアクティビティが計算リソースの乱用を示す一方、IAMアクティビティは認証情報の侵害や長期アクセスキーの不正使用、そしてクラウトサービスの不正使用の可能性を示唆しているからです。

AIインフラ保護のための重要な教訓

このインシデントの重大性は暗号通貨マイニングアクティビティそのものではなく、それが発生した場所にあります。侵害されたシステムはAmazon Bedrockサービスへのアクセス権を持つAIゲートウェイとして機能し、クラウドインフラ、アイデンティティ、そしてさまざまなAIオペレーションの交差する場所に位置していました。組織がAI機能を実運用環境に導入していくなかで、これらのプラットフォームは、露出したサービス、認証情報窃取、クラウドの設定ミスなどを通じて攻撃者がすでに狙っているアタックサーフェスの一部となりつつあるのです。

このケースでは詳細な侵入経路は特定されておらず、ワークロードの侵害と調査中に検知された疑わしいIAMアクティビティの間に決定的なつながりは確認されませんでしたが、これらのイベントは全体的な現状を裏付けています。つまり、AIインフラは個別のテクノロジースタックとして扱うのではなく、クラウド環境全体の一部として保護しなければならないとうことです。

このケースでは、最も目立った侵害の兆候は暗号通貨マイニングインフラとの通信でした。しかしここで得られたより重要な教訓は、このインシデントの全貌が理解される前にDarktraceのビヘイビア分析により明らかになった、高い権限を持つAI関連アセットを取り巻くリスクです。AIゲートウェイによりクラウド権限、モデルアクセス、アプリケーションワークフローがますます集約されるなかで、防御者は個別のアラートに集中するよりも、ワークロード、アイデンティティ、サービスの間でどのように動作がつながっているかを理解することに重点を置く必要があるでしょう。

協力:Angel Arribas Lopez (Associate Principal Cyber Analyst)、Nathaniel Jones (Field CISO/VP Threat Research)、Emma Foulger (Global Threat Ops)、Mark Turner(Security Researcher)

編集:Ryan Traill (Content Manager)

付録

Darktraceによるモデル検知結果

·       Compromise / High Priority Crypto Currency Mining

·       Compromise / Monero Mining

·       Device / Internet Facing Device with High Priority Alert

·       IaaS / Unusual Activity / Unusual AWS CLI Activity

·       IaaS / Admin / New AWS User Account Creation

MITRE ATT&CK マッピング

初期アクセス – 外部リモートサービス – T1133

初期アクセス – 有効なアカウント – T1078

実行 – コマンドおよびスクリプトインタプリタ – T1059

永続化 – アカウント作成 – T1136

探索 – クラウドサービス探索 – T1526

影響 – リソースハイジャッキング– T1496

参考資料

[1] https://docs.litellm.ai/blog/security-update-march-2026

[2] https://www.abuseipdb.com/check/145.241.123.102

[3] https://urlscan.io/search/#185.62.1.8

[4] https://www.virustotal.com/gui/file/85de36ff66fae9f4b059cbedf6d36e017ebc26c828f99f911a96e78636f21200/community

Continue reading
About the author
Angel Arribas Lopez
Associate Principal Cyber Analyst

Blog

/

AI

/

July 8, 2026

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

Default 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
あなたのデータ × DarktraceのAI
唯一無二のDarktrace AIで、ネットワークセキュリティを次の次元へ