Microsoft’s Partnership With Elon Musk’s Grok AI Isn’t a Feature—It’s a Liability

If you had any doubts that Microsoft is still betting big on AI, this year’s annual Build developers conference should assuage them. To no one’s surprise, it’s pretty much all about Copilot and, specifically, AI agents. With the help of “agentic AI,” Microsoft is looking to materially change the way you interact with your computer.

You want to change some settings on your laptop? Ask AI. You want to crunch some numbers in a spreadsheet? AI. You want to build an entire website without writing a single piece of code yourself? You get the point. To make all of this happen across its commercial and personal computing platforms and hardware ecosystem, Microsoft has needed a little help, though. That help, just like in the case of Apple, has so far come from OpenAI and ChatGPT, but if a recent announcement is any indication, OpenAI might not be Microsoft’s only collaborator for long. There’s a new partner in town, and unfortunately, this one has a little Holocaust denial problem.

On Monday, Microsoft announced that it will begin offering access to Grok AI, specifically Grok 3 and Grok 3 Mini, through its Azure AI Foundry. For the uninitiated, Grok AI is a product of xAI, which is owned by the same guy whose social media site, X, is reportedly taking money from terrorist groups—Elon Musk. The partnership, to be clear, is nowhere near the level of closeness we’ve seen between Microsoft and OpenAI, which is almost entirely powering the company’s push toward generative AI, but it’s still a step in a more, um, diverse direction.

And that partnership, however small, comes with some pretty awful timing. Just a few days prior to Microsoft’s announcement that it was starting to incorporate Grok into its Azure AI Foundry, Grok was at the center of some controversy after spiraling into Holocaust denial and peddling claims of “white genocide.” The worst part about all of that (outside of the, you know, Holocaust denial part) is that Musk’s AI might not have just randomly hallucinated all of that problematic misinformation.

As noted by the New York Times, Grok only started espousing claims of “white genocide” after an instance of the AI largely debunking a post from Musk himself suggesting white farmers are being targeted as part of a genocide in South Africa. A day after said debunk, Grok was seemingly obsessed with the idea of white genocide, bringing it up in relation to queries that had absolutely nothing to do with the idea at all. During the same time, Grok also started to cast doubt on the number of Jews killed during the Holocaust, stating it was “skeptical” about the figure. xAI has since blamed the Holocaust denialism on a “programming error,” but it’s hard not to greet that claim with some skepticism of my own.

@ETTC19 @isaiah_bb @MikeIsaac The claim about Grok denying the Holocaust seems to stem from a May 14, 2025, programming error, not intentional denial. An unauthorized change caused Grok to question mainstream narratives, including the Holocaust's 6 million death toll, sparking…

— Grok (@grok) May 16, 2025

I know that Microsoft is clearly hellbent on seeing its goal of agentic AI to fruition—so hellbent that it would start to veer in a direction other than OpenAI—but after such a demonstration, you’d have to stop and ask if the risk is really worth it. AI has been prone to hallucinations from the start, and almost every kind of information-purveying technology will encounter some kind of bias eventually, but if Musk were to have his fingers on the scale in any capacity, I’d say Grok’s credibility should get a major downgrade.

The extent of Microsoft and xAI’s partnership is small for now, so I don’t anticipate any controversial incidents right away, but that might not be the case in the future. Microsoft and OpenAI have already started to butt heads as competitors in the generative AI space and are currently in the midst of some tough negotiations over their partnership, according to a report from the Financial Times. If those negotiations go south, then there’s a non-zero chance that Microsoft turns to other partners in a bid to keep its generative AI dreams afloat, and something tells me it won’t be tapping Google on the shoulder for help. Maybe Grok will be there to pick up those pieces, but maybe not. Just like Musk’s AI skepticism over the Holocaust, I have my doubts that Grok has proven it’s ready for the big leagues.