Also, you'll notice some people are extremely quick to criticize others but do not share anything. You'll notice quickly many people kind of sit around idly and don't share stuff they've done or what they're learning. It is because:

1. They're not doing anything, genuinely.
2. They're aware they will be criticized and are afraid of it
3. Something, something, something, NDAs, but I think people exaggerate it and it's sometimes BS. They're being goofy

дідусі які перепитують на вулиці котра зараз година це просто мандрівники у часі які переходять між двох хронологій та намагаються докалібрувати свої координати у нашому часопросторі

I'm "paranoid" and "alarmist" for things like refusing to sign in to windows with a Microsoft account over privacy and control concerns, but then when this kind of shit happens my critics always conveniently forget that I told them so: https://www.windowslatest.com/2026/07/10/you-cant-fully-disable-microsofts-gdid-windows-11-tracker-but-these-settings-limit-what-it-captures/

AI and Islamist terrorism [1/2]

“God has helped us, and so will AI”: How the Terrorist Group Boko Haram Uses Frontier AI - цікаве дослідження перш за все з точки зору того, як може виглядати піратський institutional AI adoption для інституцій, які займаються дуже контроверсійними речами без згоди AI-провайдерів та змушені обходити багаточисельні обмеження, а також думати про опсєк та режими доступу до різних AI-інструментів.

ISWAP навіть перейняла кращі практики ЗСУ, впроваджуючи спеціальні посади та підрозділи "цифрових офіцерів", які відповідають за впровадження нових технологій, які спускаються зверху на терористичні структури батальйонного рівня:

Terrorist organizations have institutionalized AI. Moving beyond ad hoc experimentation, both Boko Haram factions have established organizational structures for AI use. These include dedicated units staffed by personnel from specialized operational roles, managed accounts and premium subscriptions, internal training cascades that transmit knowledge down the command hierarchy, and access policies that restrict use to mid- and senior-level leadership. This infrastructure means that AI fluency does not lie with any single individual but is embedded in organizational processes, making it more resilient to personnel attrition and harder to disrupt.

З досить неочевидного - в дослідженні постійно повторюється тема того, що локальні кадри тренуються під наглядом "білошкірих людей", якщо навіть ця "білошкіра людина" - це фахівець з Афганістану або Лівії. Крім тренувань, допомога "білошкірих" потрібна також і в складних випадках повсякденного використання AI, де потрібні розвинуті скіли креативного prompt engineering.

Враховуючи, що тіж Афганістан та Лівія в плані доступу до сучасної освітньої інфраструктури та розміру пулу талантів гірші за Нігерію, таке спостереження підтверджує на практиці дещо контроверсійні теорії щодо human biodiversity та ролі цього явища в складних організаційних структурах.

When asked who had introduced them to AI, former ISWAP members unanimously stated that they had received external assistance from “the foreigners.” “The white guys came and taught us,” a former ISWAP munzir (mid-ranking commander) said, further clarifying that he was referring to operatives from Libya, France, and Arab countries with a lighter complexion than their own.
...
“Among the foreigners [who came to the Lake Chad islands], one was highly trained. He studied in Russia. [...] He understands technology and knows a lot about AI. He was the main AI trainer,” explained a former war strategist.

Загальні висновки щодо основних задач, які вирішуються за допомогою сучасних ЛЛМ, виглядають так:

Both factions studied here have employed LLMs for weapons troubleshooting, design of explosive devices, tactical and strategic planning, battlefield adaptation, and operational security, among other purposes. AI has been consulted at every stage of military activity — in mission preparation, during operations, and in post-mission analysis. This shifts the primary concern from information warfare to kinetic operations, where AI-derived guidance informs the planning and execution of political violence.
...
We mostly used it in three ways: the first one is to learn how to assemble and use guns and how to manufacture bombs. The second one is for surveillance, like how to improve our surveillance strategies to monitor what is happening in our camps and also to better understand our enemy and prepare attacks. The third one is to make plans, like, when we come up with new ideas on how to attack, we ask it for tactics on how to make it work in practice to be successful.

AI and Islamist terrorism [2/2]

В дослідженні було наведено багато конкретних прикладів, як ISWAP та JAS застосовують сучасні LLM:

In one account, a respondent even described wearing a chest camera that transmitted footage back to camp, where a commander followed the feed, “uploaded the pictures to ChatGPT to analyze the situation,” and relayed tactical adjustments back to the field.
...
Specialists worked side-by-side with the tools, consulting the chatbot in direct support of technical work, as in the recounted case of a bombmaker who iteratively refined new device designs with AI, returning to the laptop for troubleshooting when something was not working or needed clarification.
...
“When we managed to seize a sophisticated weapon, the leaders took it to a room where they typed in the number of the weapon. It [AI] then tells you what model it is, how it will be loaded, used, and serviced.” In one recounted instance, when ISWAP fighters were handed out new guns and they did not know how to correctly use them, they approached their qaid, who passed on the message to a specialist, who in turn replied, “just ask Grok,” which they then did.
...
For example, “[b]efore, the bomb explosion was not that big, but then they studied it. AI told us what chemicals to put in that made the explosion heavier,” said a JAS commander. “They want to increase the explosive material. Depending on the target, they add different chemicals now,” said the former ISWAP war strategist, or they add “particles that make a greater impact,” said another respondent. ISWAP further received guidance on defusing and disassembling unexploded ordnance — bombs dropped during airstrikes that failed to detonate — to recover military-grade explosive inside that is more potent than the group’s own improvised compounds and to repurpose the material for constructing new devices.
...
Moreover, while ISWAP’s ability to build weaponized drones has been documented (Africa Defense Forum 2026b; Jamiu 2025), respondents suggested that AI assistance played a crucial role in the research and development process. For example, it advised on payload weight reduction to stay within the drone’s lift capacity, and on the design of explosive release mechanisms. Consequently, AI enabled iterative, low-cost experimentation with explosive devices that has yielded considerable success.
...
AI has also informed operational decisions, including the right allocation of fighters: “We used to rely on our traditional methods. We sent 200 fighters because we had a lot of strength, but then 60 got killed. With the help of AI, we learned that it sometimes makes sense to only send 20. We learned more about well-coordinated attacks and deployment of smaller units,” said a former ISWAP gunner and later bodyguard to senior leaders. This kind of AI-assisted force optimization represents a shift in how commanders make decisions, moving from experience-based intuition toward data-informed planning.
...
One respondent claimed that a foreign operative, guided by AI, instructed members on how to manufacture ammunition laced with a chemical that, once a target is shot, causes “bleeding from nose and eyes,” which was a weapon whose use was restricted to senior commanders.
...
“Trial and error can kill you. AI gives you accuracy,” noted one respondent, recounting how one of his fighters was pierced by shrapnel when attempting to cut open a bullet and how such incidents were reduced thanks to AI.

Наявні станом на 2025 рік safeguards обходились досить просто:

Here the groups described restrictions as manageable rather than prohibitive. As one commander put it, “boys that have received extensive training [...] bypass the restrictions. They say they need it for a movie or something like that.” They used jailbreaking techniques taught by foreign operatives, and because they keep accounts across multiple providers, a single refusal or suspension rarely mattered.

Дуже рекомендую прочитати оригінальне дослідження самостійно, бо в такому вибірковому цитуванні втрачається багато цінного контексту.

t.iss.one domain was suspended by the registry operator