SAN FRANCISCO — Google breached its own policies that barred use of artificial intelligence for weapons or surveillance in 2024 by helping an Israeli military contractor analyze drone video footage, a former Google employee alleged in a confidential federal whistleblower complaint reviewed by The Washington Post.
Google’s Gemini AI technology was being used by Israel’s defense apparatus at a time the company was publicly distancing itself from the country’s military after employee protests over a contract with Israel’s government, according to internal documents included in the complaint.
In July 2024, Google’s cloud-computing division received a customer support request from a person using an Israel Defense Forces email address, according to the documents included in the complaint, which was filed with the Securities and Exchange Commission in August. The name on the customer support request matches a publicly listed employee of Israeli tech firm CloudEx, which the complaint to the SEC alleges is an IDF contractor.
The request from the IDF email address asked for help making Google’s Gemini more reliable at identifying objects such as drones, armored vehicles and soldiers in aerial video footage, according to the internal documents included with the complaint. Staff in Google’s cloud unit responded by making suggestions and doing internal tests, the documents said.
At the time, Google’s public “AI principles” stated the company would not deploy AI technology in relation to weapons, or to surveillance “violating internationally accepted norms.” The whistleblower complaint alleges that the IDF contractor’s use contradicted both policies.
The complaint to the SEC alleges that Google broke securities laws because by contradicting its own publicly stated policies, which had also been included in federal filings, the company misled investors and regulators.
“Many of my projects at Google have gone through their internal AI ethics review process,” the former employee who filed the complaint said in a statement to The Post, provided on the condition of anonymity for fear of retaliation from the company. “That process is robust and as employees we are regularly reminded of how important the company’s AI Principles are. But when it came to Israel and Gaza, the opposite was true. … I filed with the SEC because I felt the company needed to be held accountable for this double standard.”
A Google spokesperson contested the whistleblower’s allegations and said the company did not violate its AI principles because the account’s usage of its AI services was too small to be “meaningful.”
“We answered a general use question, as we would for any customer, with standard, help desk information, and did not provide any further technical assistance,” a statement provided by the spokesperson said. “The ticket originated from an account with less than a couple hundred dollars of monthly spend on AI products, which makes any meaningful usage of AI impossible.”
Google documentation for its “cloud video intelligence” service says that tracking objects in video is free for the first 1,000 minutes and then costs 15 cents per minute.
A spokesperson for the SEC declined to comment. Anyone can file a complaint with the agency, which does not make them public. Complaints do not automatically lead to an investigation.
Representatives for the IDF and CloudEx did not respond to requests for comment.
In 2024, CloudEx was among the sponsors of a tech conference held south of Tel Aviv called “IT for IDF,” which featured Israeli military officials praising the importance of cloud computing to its operations in Gaza, according to the event website.
The complaint to the SEC claims that the use of Gemini described in the internal Google documents was related to Israel’s operations in Gaza without citing specific evidence. Google previously said its work for the Israeli government was “not directed at highly sensitive, classified, or military workloads relevant to weapons or intelligence services.”
Google, which used the motto “Don’t be evil” when it went public in 2004, has since its founding focused on serving consumers and businesses. More recently, company leaders have also sought defense contracts, triggering protests from some employees.
The company introduced its AI policies that barred uses related to weapons or surveillance in 2018, after employee pushback led the company to decline to renew a Pentagon contract that involved analyzing drone footage.
In 2021, Google and Amazon won a $1.2 billion cloud deal with Israel’s government known as Project Nimbus. Microsoft also provides cloud computing to the Israeli government. Amazon founder Jeff Bezos owns The Washington Post.
Some Google, Amazon and Microsoft employees have protested the companies’ work with Israel. Google fired more than 50 workers in April 2024, according to a group representing the workers, after some of them staged sit-ins at company offices and demanded it cease working with the Israeli government. Microsoft has also fired workers, after protests at its campus near Seattle.
Google in February last year updated its AI policies to remove its pledges not to apply the technology to weapons or surveillance, saying it needed to evolve to help democratically elected governments keep up in the global battle for AI dominance.
When the CloudEx employee filed the support request to Google in 2024, they described a bug that led to the Gemini AI software failing to properly analyze aerial footage some of the time. A google staffer was copied in on the initial request, according to the documents included in the complaint to the SEC. The complaint alleges that staffer works on IDF’s Google Cloud account.
After exchanging several messages with a Google support worker, the CloudEx employee said the issue resolved itself, according to the documents.
In December, the Pentagon announced that Google’s Gemini was the first AI offering to be provided to Defense Department employees on its new GenAI.mil platform, under a new drive to increase military use of the technology.
The whistleblower report adds to claims that major American tech companies have been drawn into assisting Israel in its war in Gaza, even as public opposition to the war in the U.S. grows.
Israel has said that 1,200 people, most of them civilians, died in the Oct. 7, 2023, attacks by Hamas that triggered the war. The Gaza Health Ministry, which does not differentiate between civilians and combatants, says that more than 71,000 Palestinians have been killed in the Israel-Gaza war.
Reporting from The Post and other news organizations has shown that Google and other U.S. tech giants have worked on military-related projects for Israel.
In January 2025, The Post reported that Google employees rushed to provide the Israeli military with greater access to AI tools in the weeks after the Oct. 7 attack. An internal document showed that one Google worker had warned colleagues that if requests from Israel’s Ministry of Defense for more AI capacity were not approved, the country might turn to Amazon instead.
In August, Microsoft, which also has contracts with the Israeli government, said it had opened an internal inquiry after the Guardian newspaper reported that the company’s cloud services were being used to store phone-call data obtained through large-scale surveillance of civilians in Gaza and the West Bank.
In September, Microsoft said the investigation led it to shut off a unit inside Israel’s Ministry of Defense from accessing some cloud services, in line with its terms of service barring mass survellance of civilians.
The post Google helped Israeli military contractor with AI, whistleblower alleges appeared first on Washington Post.
