Microsoft has acknowledged it has supplied AI technology to the Israel Ministry of Defense (IMOD) but has stated it “found no evidence” the technology had been used to “target or harm people in the conflict in Gaza”.
Microsoft published a statement last week, where it admitted to providing IMOD with “software, professional services, Azure cloud services, and Azure AI services, including language translation”. It added: “As with many governments around the world, we also work with the Israeli government to protect its national cyberspace against external threats”.
The statement follows reports of Israel using AI in its conflict in Gaza, which has resulted in thousands of Palestinian deaths. As reported by The Guardian last year, the Israeli military employed its own AI system, called Lavender, with intelligence sources claiming Israeli military officials permitted large numbers of Palestinian civilians to be killed.
As Microsoft has now acknowledged, there were concerns from its employees and the public regarding the use of its Azure and AI technologies by the Israeli military. In response, it has conducted an internal review, alongside an external firm the company has omitted to name.
“Based on these reviews, including interviewing dozens of employees and assessing documents, we have found no evidence to date that Microsoft’s Azure and AI technologies have been used to target or harm people in the conflict in Gaza,” the company said.
“Our relationship with the IMOD is structured as a standard commercial relationship,” it continued. “Like all our customers, the IMOD’s use of our technology is bound by Microsoft’s terms of service and conditions of use, including our Acceptable Use Policy and our AI Code of Conduct. These require customers to implement core responsible AI practices – such as human oversight and access controls – and prohibit the use of our cloud and AI services in any manner that inflicts harm on individuals or organisations or affects individuals in any way that is prohibited by law.”
Microsoft noted it does occasionally provide special access to its technologies beyond the terms of its commercial agreements. It did this in the weeks following 7th October 2023 by providing limited emergency support to the Israeli government to help rescue hostages, but with “significant oversight and on a limited basis”.
It also noted militaries “typically use their own proprietary software or applications from defense-related providers for the types of surveillance and operations that have been the subject of our employees’ questions. Microsoft has not created or provided such software or solutions to the IMOD.”
The company also acknowledged it does not have visibility into how its customers use its technology on their own servers, typically the case for “on premise software”. As such, it does not have visibility to the IMOD’s government cloud operations.
“In sum, Microsoft has long defended the cybersecurity of the State of Israel and the people who live there,” the statement concludes.
“We similarly have long been committed to other nations and people across the Middle East. Our commitment to human rights guides how we engage in complex environments and how our technology is used. We share the profound concern over the loss of civilian life in both Israel and Gaza and have supported humanitarian assistance in both places. The work we do everywhere in the world is informed and governed by our Human Rights Commitments. Based on everything we currently know, we believe Microsoft has abided by these Commitments in Israel and Gaza.”
Last year, a group of current and former Microsoft employees launched the No Azure for Apartheid petition, which currently has 1527 signatures. The group has now called for Microsoft to make this investigation public.
“It’s very clear that their intention with this statement is not to actually address their worker concerns, but rather to make a PR stunt to whitewash their image that their relationship with the Israeli military has tarnished,” said former employee Hossam Nasr to Israeli newspaper Haaretz. Nasr was fired in October for helping to organise an unauthorised vigil for Palestinians killed in Gaza at Microsoft’s headquarters.
As RockPaperShotgun reported, the international Boycott, Divestment and Sanctions movement called for a boycott of Microsoft products last month in protest against the company’s reported connections with the Israeli military. Back in February, the Associated Press published a report into the use of AI technology by the Israeli military, including Microsoft and OpenAI.
Similarly, the developer of indie role-player Tenderfoot Tactics removed the game from sale on Xbox in support of the boycott.