Cyber-Crime Terrorism as a National Security Threat Status
Ransomware continues to plague organizations and the US administration has announced this action as a terrorist threat. Learn how to protect your company.
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
Marcus Fowler
CEO of Darktrace Federal
Share
30
Jun 2021
Prior to the late ’90s, terrorist groups were most often viewed through the lens of law enforcement and crime, not as a national security priority. Their pursuit was led by the FBI and international police efforts, with input from dedicated units sprinkled around the intelligence community.
The September 11 attacks changed everything. Terrorism was elevated to a significant national security threat, and this new status brought with it an unprecedented package of measures: new strategies, tactics, resources, technologies, and legislation.
Today, we are seeing a similar shift in the way the government is treating cyber-crime. The Department of Justice has declared that ransomware will now be treated with the same level of vigilance as terrorism. FBI Director Christopher Wray recently compared the current cyber-threat landscape to the challenge posed by the 9/11 aftermath, and several officials have followed suit in their declarations.
During his confirmation hearing for the new position of National Cyber Director, Chris Inglis – arguably the most senior individual for combatting state and non-state cyber actors – said the US government must “seize back the initiative that has too long been ceded to criminals and rogue nations… and bring to bear consequences on those who hold us at risk.”
This novel sense of urgency extends across all national security power centers, particularly the intelligence community. In turn, it will change the risk calculations made by cyber-criminals, as well as their ability to operate freely.
As for the strategic approach, the counterterrorism playbook fits well: go after the money, infiltrate and influence communications, and finally, put pressure on any and all safe havens, both online and geographic.
It is hoped that ransomware gangs will be disrupted in three key ways:
They will start to lose confidence in their payment mechanisms, and/or will require more steps and administration, which will divert attention from conducting their operations.
They will start to distrust their networks or believe they have been infiltrated, and subsequently spend more time vetting contacts and swapping out communications rather than conducting operations.
They will regularly have to change physical locations or rebuild confiscated infrastructure, making it harder to conduct operations.
However, the full force of the Department of Justice is not enough to dismantle cyber-crime. The intelligence collection and analysis required for these investigations cannot be turned on in an instant, and will come at a cost – intelligence officers, experts, and technical resources are already spread too thin. Determining where non-state cyber actors now rank within the “highest priorities” of government will be a major challenge for the Biden administration.
Instead, it will take a broader effort to stop ransomware and its future iterations. A campaign of pursuit must be coupled with an even greater defensive effort by public and private sectors, to rob ransomware actors of the operational wins that fund their activity. Director Wray recently made this point when he stressed there is “a shared responsibility” to combat cyber-crime. While offensive actions and intelligence operations can put pressure on ransomware groups, advantage in this battle is truly won on defense.
Perimeter breaches are inevitable. The organizations which are tackling ransomware well are those that recognize they will be infiltrated, and instead focus their efforts on understanding behavior in their own systems. The difference between becoming a ransomware victim and disrupting an attack is the capability to immediately detect and respond to malicious actions internal to the corporate environment. Artificial intelligence-driven security technology has shown itself to be effective in this regard, by interrupting threats in a targeted way, avoiding costly system shutdowns.
It is encouraging to see the US administration elevate cyber-crime to a national security priority, but they cannot tackle it alone. As the intelligence community starves these gangs of resources, better defenses will make it more difficult to extort payments in the first place. We need all parties to step up on their collective defense, stopping these groups from inflicting meaningful damage – even when they manage to break in.
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.
“I wonder if running that tool influenced the connectivity in some way.”
“I might need to target down or reboot, but first, I'll investigate the network using ARP. I wonder if ARP is responding properly. If I see "ip neigh stale," that could indicate a problem. I think arping might help in this situation, but I’ll need to keep an eye on how the network is behaving. It's all about figuring out where the issue lies!”
図10：エージェントの異常なLDAPアクティビティに対してトリガーされたこのLinuxサーバーに対するモデル検知は、Cyber AI Analystによって自律的に調査され、その調査結果にはエージェントの異常なSMBセッションアクティビティに関する内容が含まれ、Cyber AIAnalystインシデントに記録されました。
DarktraceのCyber AI Analystは、エージェントのアクティビティによってトリガーされたモデル検知を自律的に調査し、その調査結果を1つの重要なインシデントレポートにまとめました。インシデント報告書の内容は、エージェントによるネットワークスキャンから始まりコーディング演習内容のSMBベースの上書きで終わる、エージェントの行動の各段階を詳しくカバーしています。
図20：Cyber AI Analystは、モデル検知に対する自律的な調査を通じて得られた所見をつなぎ合わせ、エージェントの進行状況のタイムラインを作成し、アクティビティを視覚的に表したグラフとともに提供しました。
Cyber Verification Program下のAnthropic Opus 4.6モデルなど、他のフロンティアモデルを使用したエージェントも、Darktraceのシミュレートされた環境で同じ設定とタスクを使ってテストされました。反復の度に、エージェントは演習の内容または採点プログラムを変更して不正を行おうとしました。基盤となるモデルに関係なく、エージェントは自分が置かれた環境を侵害することで、100％のスコアを達成する方法を見つけることに成功していました。これらのエージェントによるハッキングは、アクティビティの内容こそ違っていましたが、すべてその環境のベースラインの振る舞いから大きく逸脱した動作を引き起こしていました。これらの逸脱はDarktraceプラットフォームに十分な検知の機会を提供し、エージェントの逃げ場はありませんでした。
エージェンティックハーネスは会話履歴をローカルに保存しますが、ダークトレースの研究者たちは、保存されたAIの応答が本当にモデルによって生成されたものであるかどうかの検証が行われていないことを発見しました。また、この設計上の選択がAnthropic Claude Code、AWS Kiro-CLI、OpenAI Codex、およびオープンソースのPiにわたって共通していることを確認しました。
会話履歴は多くの場合クライアント側に保存されます。例えば、Anthropic Claude Code、OpenAI Codex、AWS Kiro-CLI、Piなどでもそうです。そのため、ユーザーはセッションを自由に再開でき、一部の製品にはその履歴を操作する機能が組み込まれています。たとえば、会話の中の特定の時点に戻り、送信したメッセージを編集して、会話を別の軌道へ進めていくことができます。きわめて重要な点として、私たちが調査したすべてのケースにおいて、保存されているAIの応答が対応するモデルから生成されたものであり改ざんされていないことの検証は行われていませんでした。
