Imagine a regular day in your life without your phone or computer, and you’ll realise how big a chunk of your routine is hung on these mobile devices.
Technology has injected unimaginable advantages into life as we know it, from connectivity to interpersonal relationships to providing any piece of information in an instant. However, such expansion of technology has resulted in insecurities and threats—like cyber frauds, cyber attacks, and so on for the countries worldwide.
Speaking at an event about the AI revolution, organised by Society to Harmonise Aspirations for Responsible Engagement (SHARE) in collaboration with Guwahati’s Cotton University, former IPS officer Bhaskar Jyoti Mahanta stressed on the AI-related developmental initiatives and their societal impacts.
“We must take heart from what DeepSeek has done,” he said. “It is the most pertinent example of how innovation and talent could trump resources.”
SHARE also released a report titled ‘Evolving Landscape of Tech-Terror Nexus and Response Options for Global South’. The paper focused on terror group activities using emerging technologies to outplay global counter-terrorism efforts. Particulars from the report are listed below:
What are Terrorism and Violent Extremists (TVEs), and how do they use AI?
In his research, Randy Borum defines ‘Terrorism and Violent Extremist (TVE)’ as an evolutionary concept that morphs based on temporal, cultural and geopolitical circumstances. Such extremist groups, operating through connected networks, are difficult to eradicate with external forces. Emerging generative AI accelerates such malpractices. A November 2023 study by Tech Against Terrorism demonstrated how AI is being used by terror outfits like al Qaeda and Daish as well as other extreme right-wing groups. With such open-source tools (like CAMeL, PyArabic and Bhashini) available for free, these groups translate their text to multiple languages without the help of any complex software.
A similar case was the hashtag hijacking of FIFA World Cup where law enforcement agencies were merely successful but the terror extremists operated it well.
TVE recruitment and communication methods and funding
AI-based technologies have been useful in recruitment of fresh cadres and ranks into terror organisations. This involves both identification of potential recruits as well as inserting the proposed ideology into their minds. The UK’s Independent Reviewer of Terrorism Legislation, Jonathan Hall was reportedly “recruited” into terrorist folds by an AI bot.
Communication is the backbone of any organisation and TVEs are no different in this regard. From the early days of al Qaeda using video tapes and emails to the human courier network used by Osama bin Laden in the post-Tora Bora days, these groups have always focused on secure communication, while also operating with chatrooms and encrypted messaging apps. Post May 2020, Sunni Islamist political organisation Hay’at Tahrir al Sham instructed its members to move away from conventional encrypted platforms like Telegram, Facebook Messenger, etc. to more advanced communication platforms like Conversations, SilentCircle, Redphone, and Signal. This move was possibly induced by the growing belief that law enforcement agencies had ways to take down groups and channels in the former platforms. Similar use of highly sophisticated apps was witnessed during the busting of scores of al Qaeda-affiliated sleeper cells in India in 2022-23.
The January 2025 issue of the Voice of Khurasan magazine (by Islamic State-Khurasan Province or IS-KP—a violent extremist organisation primarily based in Afghanistan), dedicates an entire chapter named “Light of Darkness” on various open-source encrypted messaging applications which may be leveraged by their followers.
Terrorists and violent extremists are known to using the dark web and cryptocurrencies for funding their operations. According to the US Department of Treasury, threat actors turn to the dark web to fund pages like “Fund the Islamic Struggle without Leaving a Trace.”
Evidence of cryptocurrencies being used to finance the 2015 Paris Attack and the 2021 Easter Bombing in Sri Lanka has also come to fore. AI-powered deepfakes are deployed to beat the verification processes.
Global South struggles to counter AI-enabled violent extremists
The internal security budget of countries in the Global North and South shows a clear economic gap. Wealthier nations, particularly in the G7, allocate significantly higher budgets to internal security compared to many countries in the Global South. For example, the USA's 2024 budget is estimated to be 65.9 billion USD, while India's is 25.8 billion USD. Countries like Thailand, Mexico, Egypt, Brazil, Indonesia, Argentina and South Africa were not assigned any such budget. This difference makes it harder for law enforcement agencies in less prosperous nations to tackle security challenges effectively.
Unavailability of data in digitised format and the lack of access to large datasets within these law enforcement agencies prevent proactive actions.
Tackling tech terror, especially in Global South languages
The first step is to ensure flawless, machine-readable data which can be pre-processed to mitigate data-entry errors. This requires restructuring data and involves four steps: data cleaning, data integration, data transformation and data reduction. This is critical especially when multi-source input mechanisms are used to generate the master database.
A secure database is crucial, protected by TrustTech solutions like Fully Homomorphic Encryption (FHE) and Searchable Symmetric Encryption (SSE), used in sensitive US databases. While developing the data lake, the use of TENSOR (clusTEriNg terroriSm actiOn pRediction) framework, which enables multi-source data collection, is advised.
Traditional content moderation techniques of various platforms are generally bypassed using local languages and dialects. To overcome this challenge, the automated crawler bots deployed for the collection of OSINT needs to be integrated with various open-source language transliteration services. Thus, the same AI-based tools used by the TVEs could be adopted to hunt them down.
The CNN-LSTM hybrid model has shown high accuracy (96–99.2%) in predicting terrorist incidents. This AI-powered system combines convolutional neural networks (CNN) for analyzing local data patterns and long short-term memory (LSTM) networks for extracting context-based features.
