From Propaganda to Attack Planning: How Terrorists Use Neural Networks

cryptonews.ru2026-08-07 tarihinde yayınlandı2026-08-07 tarihinde güncellendi

Özet

Terrorist organizations are increasingly using artificial intelligence for propaganda, recruitment, security, and operational planning, according to a July 2026 CSIS report. The analysis covers networks linked to al-Qaeda* and the Islamic State* (IS). While traditional weapons like firearms and explosives still account for over 80% of attacks, AI is transforming how extremist structures prepare and promote their activities. Their strategies differ: IS focuses on speed and immediate impact, using AI for content like synthetic presenter videos since 2023. Al-Qaeda affiliates, prioritizing long-term resilience, have used AI-generated imagery and AI-powered translation tools to reach multilingual audiences. Research involving former members of Boko Haram* and its offshoot ISWAP* revealed they sought operational support from AI systems like ChatGPT, Claude, and Gemini in 2025-2026. During that period, ISWAP dramatically increased its attacks in Nigeria. AI lowers organizational costs, enabling smaller networks and lone actors to perform tasks once requiring significant resources. While states gain advanced surveillance tools, terrorists exploit open-source models and bypass built-in safeguards. The technology's spread continues to shift the balance between state monitoring capabilities and the autonomous action capacity of small groups. The pattern mirrors the early adoption of the internet by extremists. A key technical paradox is that restrictions on powerful AI models by ma...

Terrorist organizations are increasingly using artificial intelligence for propaganda, recruitment, security, and operational planning — these are the conclusions of the Center for Strategic and International Studies (CSIS) report "Artificial Intelligence and the Future of Terrorism," published on July 10, 2026, by researchers Daniel Byman and V.S. Subrahmanian. The work covers networks associated with Al-Qaeda* and the Islamic State** (IS).

According to the same report, firearms and various types of explosive devices accounted for over 80% of all terrorist attacks in the past decade. Technology is not yet replacing traditional methods but is noticeably changing how extremist structures prepare and promote their activities.

Different Strategies of IS and Al-Qaeda

The approach of the Islamic State and Al-Qaeda to artificial intelligence differs noticeably and corresponds to the organizational culture of each group: the former focuses on speed and immediate effect, the latter on long-term sustainability.

Supporters of the Islamic State have been creating content using AI since 2023, including videos with synthetic presenters. In June 2025, the English-language issue of the Voice of Khorasan magazine, associated with ISIS-Khorasan*** and published by Al-Azaim Foundation, called proficiency in AI technology an individual duty for every supporter.

Supporters of Al-Qaeda have been disseminating AI-generated images also since 2023. The Al-Shabaab**** affiliated group, meanwhile, used AI-based translation tools to reach audiences in different languages — this was reported by the UN Monitoring Group in report S/2026/44, covering the period up to December 2025.

Boko Haram and Chatbots

Cambridge University researcher Antonia Julich conducted 57 interviews with 27 former members of the Boko Haram***** group, including ISWAP******, in 2025–2026. Former militants reported turning to AI systems for operational support, among them:

  • ChatGPT

  • Claude

  • Gemini

  • Grok

  • Meta AI

  • DeepSeek

The researcher's report "God has helped us, and so will AI" was released in July 2026, and her fresh comments on the topic appeared on August 3, 2026.

According to the Global Terrorism Index 2026, based on statistics for 2025, ISWAP in Nigeria increased the number of attacks from 20 to 92 over a year, resulting in 384 deaths.

Costs Fall, Control Becomes More Complex

Artificial intelligence reduces organizational costs: small networks and individuals gain the ability to perform tasks that previously required significant resources — creating propaganda, maintaining contact with potential supporters, operational support. States, meanwhile, obtain more sophisticated surveillance and analysis tools; however, terrorist structures use open models and find ways to bypass built-in restrictions.

The totality of this data shows that artificial intelligence has become part of the daily practice of extremist groups of various profiles — from large international networks to regional offshoots like ISWAP. The further spread of technology will continue to change the balance between states' monitoring capabilities and the ability of small groups to act autonomously.

AI Opinion

From the perspective of machine data analysis, the situation repeats a pattern from two decades ago, when extremist networks were the first to master internet forums and social media — long before states developed monitoring mechanisms. Technology each time outpaces regulatory response by several years.

The article does not address a separate technical aspect: the AI industry has already split into closed models with enhanced protection and open alternatives without such restrictions. Major developers, including Anthropic and OpenAI, restrict access to their most powerful cybersecurity models for the general audience, as previously reported by Hash Telegraph. The paradox is that such restrictions protect legitimate users but have little effect on groups that rely on open models and find ways to bypass filters. Will this access gap remain a deterrent or turn into another arms race between defenders and attackers?


* The organization is recognized as terrorist, its activities are prohibited in Russia.
** The organization is recognized as terrorist, its activities are prohibited in Russia.
*** Is a structural unit of an organization recognized as terrorist and banned in Russia.
**** The organization is recognized as terrorist, its activities are prohibited in Russia.
***** The organization is recognized as terrorist, its activities are prohibited in Russia.
****** Is a structural unit of/is linked to an organization recognized as terrorist and banned in Russia.

end-content

İlgili Sorular

QAccording to the CSIS report cited in the article, what are the main ways terrorist organizations are using artificial intelligence?

AAccording to the CSIS report, terrorist organizations are increasingly using artificial intelligence for propaganda, recruitment, operational security, and planning attacks.

QHow do the approaches of Islamic State (IS) and Al-Qaeda to using AI differ, as described in the article?

ATheir approaches differ according to their organizational culture: Islamic State focuses on speed and immediate impact, using AI to create content like synthetic video presenters. Al-Qaeda focuses on long-term resilience, using AI for tasks like generating images and, through affiliates like Al-Shabaab, for AI-powered translation tools to reach audiences in different languages.

QWhat specific AI systems did former Boko Haram/ISWAP militants report using for operational support in the Cambridge University study?

AFormer militants reported using AI systems including ChatGPT, Claude, Gemini, Grok, Meta AI, and DeepSeek for operational support.

QWhat key trend does the article highlight regarding the impact of AI on the operational capabilities of terrorist groups?

AThe article highlights that AI is lowering organizational costs, enabling smaller networks and lone actors to perform tasks that previously required significant resources, such as creating propaganda and providing operational support, thus increasing their autonomy.

QWhat paradox related to AI model access and security does the article's 'AI Opinion' section mention?

AThe paradox is that while major developers restrict access to their most powerful, security-hardened AI models to protect legitimate users, these restrictions have little effect on terrorist groups who rely on open models and find ways to bypass filters. This creates a gap that may become a new arms race between defenders and attackers.

İlgili Okumalar

İşlemler

Spot
活动图片