At the May 19 Microsoft Build developer conference in Seattle, artificial intelligence (AI) security chief Neta Haiby accidentally disclosed confidential details about Walmart’s AI plans. The leak happened during a Tuesday panel after Haiby shared her screen in a live session, inadvertently exposing an internal Teams conversation discussing Walmart’s AI integration.
As first reported by tech news publication the Verge, the conversation was posted by principal cloud solution architect Leigh Samons. It revealed that Walmart was preparing to adopt Microsoft’s Entra Web and AI Gateway, with Samons stating the retail giant was “ready to ROCK AND ROLL” with the tools.
Walmart will reportedly embed Microsoft’s AI technologies into Walmart’s operations through a tool called “MyAssistant.” It was built by Walmart using proprietary data and Microsoft’s Azure OpenAI Service, first introduced last summer to support store associates in summarizing documents and generating marketing content.
The Teams chat also referenced an unnamed senior Walmart AI engineer applauding Microsoft’s progress in AI safety, stating, “Microsoft is WAY ahead of Google with AI Security. We are excited to go down this path with you.”
Protest disrupts session, exposes internal chat
The leak occurred after protesters heckled attendees of the session, forcing Haiby to change her screen share, accidentally displaying the confidential exchange to a room full of attendees. According to a Tumblr post reviewed by CNBC, Haiby, who previously served in the Israeli Defense Forces, was on stage alongside Sarah Bird, Microsoft’s head of responsible AI.
Hossam Nasr, a member of the activist group “No Azure for Apartheid,” confronted Bird during the presentation, saying, “Sarah Bird, you are whitewashing the crimes of Microsoft in Palestine.” The livestream audio was muted as Nasr continued to speak.
In the early hours of Tuesday, another Palestinian tech worker interrupted Jay Parikh, Microsoft’s head of CoreAI. The protesters, who asked reporters to keep their identity anonymous out of fear of retaliation, accused Parikh of complicity in Israeli military actions in Gaza.
Microsoft’s internal disagreements over Israel ties continue
The disruptions at Build are part of several cases of internal dissent and public protest at Microsoft events. At the company’s 50th anniversary celebration in April, two engineers interrupted presentations to denounce Microsoft’s involvement with Israel. Both were dismissed shortly thereafter, according to internal documents obtained by CNBC.
During that April event, Ibtihal Aboussad, a software engineer in Microsoft’s AI division, barged in on Microsoft AI CEO Mustafa Suleyman’s speech.
“Mustafa, shame on you,” she said, walking toward the stage. “You claim that you care about using AI for good, but Microsoft sells AI weapons to the Israeli military. Fifty thousand people have died, and Microsoft powers this genocide in our region.”
AI’s presence in military operations
Several news sources in January talked about leaked documents from Israel’s Defense Ministry and Microsoft’s Israeli subsidiary, which showed that Microsoft has a presence in all major branches of Israel’s military infrastructure.
According to internal commercial records, sales of Microsoft’s cloud and AI services to the Israeli military have surged since the start of the Gaza offensive in October 2024.
The documents show that multiple Israeli army units, including air, ground, naval forces, and intelligence Unit 8200, are using Microsoft’s Azure cloud platform. The military also had access to OpenAI’s GPT-4 language model through Microsoft’s partnership with the AI company.
However, the company released a statement on May 15, arguing there is “no proof that Israel’s military used the company’s AI tools to harm civilians in Gaza.”
Meanwhile, several AI companies have now partnered with defense agencies, walking back on the basis that they were “against military use of their technologies.”
