Google signed off on Project Nimbus despite warnings of misuse
5 Min Read
Google proceeded with Project Nimbus deal despite internal warnings that it cannot monitor or prevent “Israel” from potentially using its cloud and AI technology to commit human rights violations against Palestinians.
-
A person rides past the Google sign outside the Google offices in Sunnyvale, California, April 18, 2024. (AP/Terry Chea)
In a deeply revealing investigation published by The Intercept, journalist Sam Biddle lays out how Google knowingly proceeded with its Project Nimbus contract with “Israel” despite serious internal warnings about the potential misuse of its cloud and AI technologies in ways that could contribute to human rights abuses against Palestinians.
According to internal Google documents obtained by The Intercept, the tech giant understood that it would have “very limited visibility” over how “Israel”, including its military and intelligence agencies, would use its cloud services. Legal experts interviewed by Biddle suggested this foreknowledge could expose Google to potential liability under international law.
Quoting international law expert León Castellanos-Jankiewicz, Biddle writes, “They’re aware of the risk that their products might be used for rights violations. At the same time, they will have limited ability to identify and ultimately mitigate these risks.”
This legal and ethical alarm bell was raised before Google and Amazon jointly secured the multi-billion-dollar Project Nimbus deal with the Israeli occupation in 2021. The internal report, codenamed “Selenite” at Google, was part of a due diligence assessment. It openly warned that Google Cloud services could be linked to “facilitation of human rights violations, including Israeli activity in the West Bank,” which would result in “reputation harm” and complicity concerns.
Collaboration with Israeli security agencies
One of the most troubling aspects revealed by Biddle is that the Project Nimbus agreement includes direct collaboration with “Israel’s” security establishment, such as the Ministry of Security and Shin Bet. The report details the formation of a Classified Team within Google composed of Israeli nationals with clearances, tasked with handling sensitive workloads that cannot be shared with the wider company.
According to The Intercept, this level of collaboration, including joint drills and intelligence sharing, is “unprecedented in Google’s deals with other nations.”
A third-party review conducted by Business for Social Responsibility (BSR) had advised Google to withhold machine learning and AI tools from “Israel”, given the heightened risk of misuse. Yet The Intercept notes that those technologies were reportedly made available, despite contractual limitations that essentially shield the Israeli regime from external scrutiny.
No oversight, no exit
Biddle highlights that the Nimbus contract limits Google’s ability to monitor how its services are used and restricts its legal obligations to act, even in the face of rights abuses. If another country’s legal system were to investigate “Israel’s” use of the tech, Google would be contractually obligated to “Reject, Appeal, and Resist” foreign legal requests, notify “Israel” immediately, and potentially violate international legal orders in the process.
Crucially, according to documents reviewed by The Intercept, “Israel” can extend the deal for up to 23 years, and the terms prevent Google from unilaterally terminating the contract. Google Cloud CEO Thomas Kurian personally signed off on the deal, fully aware of these implications.
Legal exposure and corporate precedent
Biddle draws on legal precedent to explore whether Google might face consequences. He cites cases involving IBM, Lafarge, and companies tied to apartheid South Africa, though such accountability is rare. While corporations aren’t legally bound under international human rights frameworks, civil suits or investigations by bodies like the ICC could target individual executives if software is shown to enable war crimes or crimes against humanity.
Castellanos-Jankiewicz explains, “Google knew or should have known that this software was going to be used or is being used in the commission of serious human rights violations.”
Alphabet’s shareholders push back
Despite shareholder efforts to compel Alphabet to conduct an independent audit of human rights risks linked to Project Nimbus, Google’s board opposed the proposal. As Biddle reports, the board insisted that its existing processes, including Google’s AI Principles and its commitment to the UN Guiding Principles on Business and Human Rights, were sufficient.
Yet critics argue that these principles ring hollow if the company deliberately avoids implementing due diligence. “It sounds like Google is giving the Israeli military a blank check to basically use their technology for whatever they want,” Castellanos-Jankiewicz tells The Intercept.
Biddle closes by pointing out that Project Nimbus was projected to generate over $3.3 billion between 2023 and 2027, not just from “Israel’s” military but also its corporate and financial sectors. In that light, Google’s internal warnings, and its decision to move forward anyway, offer a sobering insight into the tech industry’s willingness to compromise oversight for revenue, even in conflict zones widely described as sites of ongoing war crimes.
