In Q4 2023, Israel’s Unit 8200 developed a ChatGPT-style AI and surveillance tool to analyze intercepted Palestinian communications, creating faster identification, intelligence reporting, and arrests, according to The Guardian, +962 Magazine, and Local Call.
The Israeli intelligence tool, trained on Arabic conversations gathered through telecom infiltration, – happened before October 7 – will lead to increased detentions of Palestinians by the occupying power.
The in-depth investigation by The Guardian, +962 Magazine, and Local Call, revealed the tool’s usage of Large Language Models (LLMs) to translate, analyze, and generate text from intercepted data, even though the Israeli intelligence 8200 Unit alleged it’s targeting “troublemakers.”
Critics, however, disagree with the Israeli Unit’s AI in military intelligence pretenses on the tool’s usage, warning the technology will lead to even more widespread surveillance and unjust detentions, further fueling tensions in the Middle East.
The Unit 8200’s purpose behind the AI in military surveillance tool was to analyze intercepted communications so that Israeli soldiers can identify individuals, create intelligence reports, and accelerate the charges and arrest process.
“AI magnifies power,” an intelligence source had stated to +972 Magazine, adding “It allows the system to process and act on the data of far more people than before. With a tool like this, I can track human rights activists, I can monitor Palestinian construction in Area C, I can see who is speaking to whom. When you possess this level of intelligence, you can manipulate it for any purpose you choose.”
ChatGPT or AI and Surveillance Tool?
The scale of AI and surveillance in this model is huge, similar to other tools like Lavendar which also infiltrated Meta’s WhatsApp groups, Where’s Daddy, Gospel and many more developed by Unit 8200.
Sources indicate that Unit 8200’s AI surveillance technology has learned 100 billion words of spoken Arabic, a large majority of which consists of intercepted calls, texts, and other digital communications.
The AI surveillance system was spearheaded by Israeli intelligence officer Chaked Roger Joseph Sayedoff, who introduced himself as the project head. At the DefenseML conference in Tel Aviv, Sayedoff said the military’s goal was to build the “largest dataset possible” of Arabic-language content.
Sayedoff acknowledged the challenges with universal language models with Arabic dialects, emphasizing – from Israel’s POV – the need for Israel to develop its own LLM “based on the dialects that hate us.”
Human rights activists argue that it is a severe violation of privacy and digital rights, at its worst level.
Zach Campbell, a senior researcher at Human Rights Watch, termed it as “highly personal information, taken from people who are not suspected of any crime, to train a tool that could later be used to justify their arrest.”
Not the First Time…
Very similar to ChatGPT, which is trained on publicly available web data, the Israeli AI mass surveillance model is fueled by a database obtained through infiltrating telecom networks, used to monitor Palestinian civilians.
In response to the war, Unit 8200 recruited Israeli citizens with machine learning expertise from tech giants like Google, Meta, and Microsoft to enhance its language models. An Israeli security source indicated that the contribution of such tech specialists was the introduction of abilities “previously accessible only to a very exclusive group of companies worldwide.”
While Google confirmed that a number of its employees are part of military reserve units, it added that their work is separate from company operations.
Meta and Microsoft refused to comment.
With another trial trying to refine the AI for surveillance tool, Unit 8200 faced problems when OpenAI refused to let the Israeli military access its LLM, preventing its use in Israel’s intelligence systems. Military investigation by +972 Magazine and Local Call uncovered that the military was able to access OpenAI’s model via Microsoft Azure.
This indicates one thing and one thing only, US tech companies can be exploited for malicious purposes, pushing users to constantly be on high alert about how factual the statements about Chinese tech are. As tensions are heightening between the US and is neighboring country Canada this could mean a lot for the tech world.
Could the tech world be divided US and Israel vs Canada and China? Canada can still cooperate with China on technological innovations because of its economic interest, especially on sectors such as AI, clean energy, and quantum computing, in which China acts as a world leader.
Canada, however, can balance well between China and the US since its strategic balancing approach will allow for taking opportunities and managing geopolitical risks and AI surveillance ethics.
Inside Telecom provides you with an extensive list of content covering all aspects of the tech industry. Keep an eye on our Intelligent Tech sections to stay informed and up-to-date with our daily articles.