Earlier this week we reported Google's $45 million Hasabra deal with the Israeli Military which has been widely criticised for ethical reasons.

Today, it's been reported that Microsoft terminated Israeli government access to cloud and AI services after a report alleged the technology was being used for mass surveillance of Palestinians. Israeli officials have yet to address the allegations.

According to the report, Microsoft's technology was being used by Unit 8200, Israel’s most powerful military intelligence unit. Credible reports have suggested that 8200 played a significant role in the development and staging of the 2024 exploding pager operation in Lebanon.

Several other Israeli military departments have also been implicated in this report, including the Ofek Unit, which manages an AI-assisted targeting program; the Matspen Unit, which builds combat systems; the Sapir Unit, which runs intelligence IT infrastructure; and the Military Advocate General's Corps, which prosecutes Palestinians and soldiers.

In early 2025, a Microsoft security team flagged suspicious activity within their defense account, prompting an internal investigation which ultimately concluded that Israel's use of Azure (cloud software) had serious ethical and legal implications.

"This was not a narrow intelligence program but a bulk collection system inconsistent with Microsoft's contractual terms," said one Microsoft investigator. 

The system stored intercepted calls and related metadata in a structured database on Microsoft’s U.S.-based servers. This was directly in violation of Microsoft's terms of service, which prohibits bulk civilian surveillance data being warehoused on its infrastructure. 

The surveillance program seemed to be intercepting data from Gaza and the West Bank which was then funnelled into AI systems for analysis. Mass civilian data being intercepted and used for monitoring purposes oversteps many ethical boundaries. It is a violation of privacy to monitor civilians who are, if anything, victims of the conflict and not combatants.

"This was not a localized system," an internal review noted. "The IDF designed it to scale across regions, and in doing so, it directly contravened contractual limits."

It is important to note, that if breached, this system has the possibility to expose not just Palestinian civilian data but also the wider Azure networks— systems which also house data from the U.S. government and many publicly traded companies.

Americans living in Israel, aid workers in Gaza, and even U.S. officials and journalists in contact with Palestinians may have had their phones calls or metadata swept into the IDF’s system. Because this database was hosted on Azure’s U.S. based servers, this content was effectively processed within American jurisdiction, raising imperative questions of legality and sovereignty.

Israel’s 8200 unit has an extensive history of cooperating with NASA and other western agencies and is plugged into data flows do cross U.S. networks.

Should further investigation be conducted to determine if U.S. government or U.S. civilians were targeted by this mass surveillance operation?