T1588.007: Artificial Intelligence
Adversaries may obtain access to generative artificial intelligence tools, such as large language models (LLMs), to aid various techniques during targeting. These tools may be used to inform, bolster, and enable a variety of malicious tasks, including conducting Reconnaissance, creating basic scripts, assisting social engineering, and even developing payloads.[1]
For example, by utilizing a publicly available LLM an adversary is essentially outsourcing or automating certain tasks to the tool. Using AI, the adversary may draft and generate content in a variety of written languages to be used in Phishing/Phishing for Information campaigns. The same publicly available tool may further enable vulnerability or other offensive research supporting Develop Capabilities. AI tools may also automate technical tasks by generating, refining, or otherwise enhancing (e.g., Obfuscated Files or Information) malicious scripts and payloads.[2] Finally, AI-generated text, images, audio, and video may be used for fraud, Impersonation, and other malicious activities.[3][4][5]
Analyst context for executives and security teams
This technique matters because adversaries can use generative AI before an intrusion to scale preparation: drafting believable phishing or vishing content, supporting reconnaissance, assisting vulnerability research, and refining scripts or payloads. For leaders, the risk is less “AI as a standalone attack” and more AI lowering the cost, speed, language quality, and realism of familiar pre-compromise activity such as fraud, impersonation, and phishing.
Executive priority
Prioritize this as a resilience and readiness issue for pre-compromise defense. Executives should ask whether the organization can recognize AI-enabled social engineering, validate high-risk requests, preserve evidence from suspected phishing/vishing or fraud attempts, and connect threat intelligence to incident response decisions. The ATT&CK relationships also tie this behavior to resource development and named campaigns/groups/software, so it should inform control prioritization without assuming local exposure or guaranteed detection.
Technical view
T1588.007 is a PRE-platform, Resource Development sub-technique under Obtain Capabilities. There is no official ATT&CK detection text, but DET0842 is listed as a related detection strategy and M1056 Pre-compromise is listed as mitigation. SOC, IR, and detection teams should validate indirect coverage around AI-assisted phishing, phishing for information, impersonation, vishing, suspicious script/payload refinement indicators, and pre-compromise threat intelligence. Treat detections as contextual: the observable event is usually the phishing, impersonation, fraud, reconnaissance, or payload behavior, not “AI use” itself.
Likely telemetry
- Email and messaging security logs related to phishing or phishing-for-information attempts
- User-reported suspicious communications, vishing, impersonation, and fraud attempts
- Help desk, finance, executive support, and security intake records for unusual high-risk requests
- Threat intelligence reporting on adversary resource development and AI-enabled preparation
- Web, social, and brand monitoring evidence for impersonation or fraudulent content where collected
Detection direction
- Do not rely on detecting AI-generated content alone; validate the downstream behaviors described by ATT&CK, especially phishing, impersonation, vishing, reconnaissance support, and script or payload enhancement.
- Tune workflows so user reports, fraud signals, and SOC alerts can be correlated during pre-compromise activity rather than handled as isolated events.
- Review DET0842 in local ATT&CK mappings and confirm what evidence sources it depends on, since the object itself does not provide official detection logic.
- Account for false positives: legitimate use of generative AI by employees, marketing teams, developers, or support staff can resemble some content-generation patterns without being malicious.
- Use relationship context cautiously: ATT&CK links this technique to campaigns, groups, and LazyWiper, but local detection should still be based on observed behavior and environment evidence.
Mitigation priorities
- Start with M1056 Pre-compromise priorities: reduce exposed attack surface, identify adversarial preparation efforts, and make targeting harder before an intrusion begins.
- Strengthen business-process verification for sensitive requests that could be enabled by AI-generated text, voice, images, or video impersonation.
- Ensure incident response playbooks cover phishing, vishing, fraud, and impersonation evidence preservation, not only endpoint compromise.
- Use threat intelligence to brief executive support, finance, help desk, SOC, and IR teams on realistic AI-enabled social engineering scenarios.
- Review public-facing information and communication channels that could support reconnaissance or impersonation, then reduce unnecessary exposure where feasible.
Analyst notes and limits
The decision value is in treating AI as an accelerator for known ATT&CK behaviors rather than a separate magic capability. Coverage depends on whether the organization can observe pre-compromise signals, user reports, suspicious communications, and downstream payload or credential-targeting activity. The related software LazyWiper is Windows PowerShell, but this technique’s own platform is PRE, so platform-specific assumptions should not be generalized from that relationship.
ATT&CK provides no official detection text for this object. The supplied mitigation relationship is high-level and partially truncated. The object supports that adversaries may obtain AI tools and that related campaigns/groups/software use the technique, but it does not prove activity against any specific organization or guarantee that AI use is directly observable.
Artificial Intelligence
Adversaries may obtain access to generative artificial intelligence tools, such as large language models (LLMs), to aid various techniques during targeting. These tools may be used to inform, bolster, and enable a variety of malicious tasks, including conducting Reconnaissance, creating basic scripts, assisting social engineering, and even developing payloads.[1]
For example, by utilizing a publicly available LLM an adversary is essentially outsourcing or automating certain tasks to the tool. Using AI, the adversary may draft and generate content in a variety of written languages to be used in Phishing/Phishing for Information campaigns. The same publicly available tool may further enable vulnerability or other offensive research supporting Develop Capabilities. AI tools may also automate technical tasks by generating, refining, or otherwise enhancing (e.g., Obfuscated Files or Information) malicious scripts and payloads.[2] Finally, AI-generated text, images, audio, and video may be used for fraud, Impersonation, and other malicious activities.[3][4][5]
How security teams should use this page
Treat this object as behavior context, not an attribution claim. Validate the related groups, software, data sources, and mitigations against official ATT&CK relationships and your own telemetry before making control-coverage decisions.
Related techniques
This mirrors the MITRE pattern of making group, software, campaign, and technique relationships scannable. Relationship notes come from mirrored ATT&CK relationship text when available.
| Domain | ID | Name | Relationship / procedure |
|---|---|---|---|
| Enterprise | T1588 | Obtain Capabilities | This object subtechnique of Obtain Capabilities. |
Groups, software, and campaigns
G1052: Contagious Interview
Contagious Interview is a North Korea–aligned threat group active since 2023. The group conducts both cyberespionage and financially motivated operations, including the theft of cryptocurrency and user credentials. Contagious Interview targets Windows, Linux, and macOS systems, with a particular focus on individuals engaged in software development and cryptocurrency-related activities. [1][2][3][4][5][6][7][8]
G0007: APT28
APT28 is a threat group that has been attributed to Russia's General Staff Main Intelligence Directorate (GRU) 85th Main Special Service Center (GTsSS) military unit 26165.[1][2] This group has been active since at least 2004.[3][4][5][6][7][8][9][10][11][12][13]
APT28 reportedly compromised the Hillary Clinton campaign, the Democratic National Committee, and the Democratic Congressional Campaign Committee in 2016 in an attempt to interfere with the U.S. presidential election.[5] In 2018, the US indicted five GRU Unit 26165 officers associated with APT28 for cyber operations (including close-access operations) conducted between 2014 and 2018 against the World Anti-Doping Agency (WADA), the US Anti-Doping Agency, a US nuclear facility, the Organization for the Prohibition of Chemical Weapons (OPCW), the Spiez Swiss Chemicals Laboratory, and other organizations.[14] Some of these were conducted with the assistance of GRU Unit 74455, which is also referred to as Sandworm Team.
S9039: LazyWiper
LazyWiper is a destructive malware observed targeting a manufacturing sector company during the 2025 Poland Wiper Attacks. LazyWiper is a native Windows PowerShell script that is believed to have been generated by a large language model (LLM). LazyWiper overwrites files on the system using the C# function `WriteRandomBytes()` and can target multiple specific file types by their extensions.[1]
C0062: Anthropic AI-orchestrated Campaign
The Anthropic AI-orchestrated Campaign was conducted in September 2025 by a likely China nexus espionage actor identified as GTG-1002. The Anthropic AI-orchestrated Campaign was a highly coordinated operation that manipulated Claude Code to perform reconnaissance, vulnerability discovery, exploitation, lateral movement, credential harvesting, data analysis, and exfiltration operations at approximately 30 entities in the technology, financial, chemical, and government sectors. During the Anthropic AI-orchestrated Campaign, human operators used Claude Code agents and Model Context Protocol (MCP) tools to automate cyber operations. Operators broke attacks into discrete tasks, used crafted prompts, and established personas to bypass AI guardrails, enabling the agents to execute the operations with minimal human involvement.[1][2]
C0063: 2025 Poland Wiper Attacks
2025 Poland Wiper Attacks is a Russian state-sponsored campaign that conducted destructive cyberattacks against Polish energy infrastructure in December 2025. Targets included more than 30 wind and photovoltaic farms, a combined heat and power (CHP) plant, and a manufacturing sector company. The attacks on the distributed energy resources (DER) disrupted communications between affected facilities and the distribution system operator, but did not impact electricity generation or heat supply. Across the campaign, threat actors deployed two previously undocumented wiper tools, DynoWiper, a Windows-based wiper and LazyWiper, a PowerShell wiper, distributed via malicious Group Policy Objects. At the CHP plant, threat actors had maintained access since at least March 2025, using that foothold to obtain credentials and move laterally before attempting wiper deployment. Some reporting has assessed the activity to be consistent with Russian Federal Security Service (FSB) threat activity group Dragonfly, also tracked as STATIC TUNDRA, while other reporting attributes the destructive wiper activities to the Russian General Staff Main Intelligence Directorate (GRU) threat activity group ELECTRUM, also tracked as Sandworm Team.[1][2][3][4]
All related ATT&CK context
Mitigation direction
Object version and sync metadata
The fields below describe the current mirrored snapshot. When Glexia retains multiple ATT&CK source imports, you can open the table to compare the same object across releases (hashes and MITRE timestamps). For MITRE’s own release notes and roadmap, see ATT&CK resources — Updates .
Imported snapshots across ATT&CK releases (1)
| Release | Bundle imported | Object version | Modified | Status | Raw hash |
|---|---|---|---|---|---|
| 19.1 | 1.1 | Current bundle | 10ae5322fd40… |
Mirrored ATT&CK source object
The raw object is retained through the mirrored ATT&CK source bundle and object hash. The raw endpoint returns the exact object from the mirrored bundle when available.
External references and citations
MITRE external references are preserved separately from Glexia analysis so citations remain traceable to their original source records.
-
[1]
MSFT-AI
Microsoft Threat Intelligence. (2024, February 14). Staying ahead of threat actors in the age of AI. Retrieved March 11, 2024.
Open source URL -
[2]
OpenAI-CTI
OpenAI. (2024, February 14). Disrupting malicious uses of AI by state-affiliated threat actors. Retrieved September 12, 2024.
Open source URL -
[3]
Google-Vishing24
Emily Astranova, Pascal Issa. (2024, July 23). Whose Voice Is It Anyway? AI-Powered Voice Spoofing for Next-Gen Vishing Attacks. Retrieved March 18, 2025.
Open source URL -
[4]
IC3-AI24
IC3. (2024, December 3). Criminals Use Generative Artificial Intelligence to Facilitate Financial Fraud. Retrieved March 18, 2025.
Open source URL -
[5]
WSJ-Vishing-AI24
Catherine Stupp. (2019, August 30). Fraudsters Used AI to Mimic CEO’s Voice in Unusual Cybercrime Case. Retrieved March 18, 2025.
Open source URL -
[6]
mitre-attack T1588.007Open source URL
Source: MITRE ATT&CK®. © 2026 The MITRE Corporation. This work is reproduced and distributed with the permission of The MITRE Corporation. MITRE ATT&CK and ATT&CK are registered trademarks of The MITRE Corporation. Glexia is not affiliated with or endorsed by MITRE.