As Elon Musk’s Grok chatbot faces global scrutiny for generating sexualized deepfakes and calling itself “MechaHitler,” a former Grok executive now serving as a “chief AI officer” at the Department of Commerce has been granted a “highly unusual” waiver to keep his stake in Grok’s parent company as he helps craft industry policy.

In November, Robert Hayes, a senior adviser and acting chief AI officer at the US Patent and Trademark Office, was granted a rare exemption from federal conflict-of-interest law to keep his 22,100 shares in xAI, the artificial intelligence company behind Grok, per ethics records obtained by the Lever.

In his role, which he began in September, Hayes is responsible for crafting intellectual property policy to “advance a positive future for” artificial intelligence. Under the Biden administration, the Department of Commerce and the US Patent and Trademark Office were tasked with studying and mitigating the harms of “deepfake” artificial images and videos.

Hayes was a “direct report” to Musk while working on Grok, according to federal ethics records, and “spearheaded xAI’s revenue and operations transformation.” Yet while his ongoing financial stake in the company would normally violate federal ethics law, the US Patent and Trademark Office offered Hayes a broad exemption, claiming he was “unable to fully divest [his] interests,” given xAI’s restrictions on share transfers.

As of October 31, Hayes’s shares in Grok were estimated to be worth less than $1 million — but that was before xAI raised $20 billion in a funding round this week.

Richard Painter, a law professor at the University of Minnesota and a former chief White House ethics lawyer, called the carveout “highly unusual.” Not only will the exemption allow Hayes to keep his equity in xAI, Painter explained, but it also allows him to participate in matters that impact the AI industry at large.

“These types of waivers are very rare for senior officials,” Painter wrote in an email to the* Lever*.

The US Patent and Trademark Office declined to comment. Hayes did not respond to additional inquiries.

This is just the latest victory that Grok has notched under President Donald Trump’s administration. In July, the Department of Defense inked a $200 million deal with xAI, and on December 22, the Pentagon announced it would work with the company to launch “the War Department’s bespoke AI platform,” integrating “the Grok family of models” across the military.

In Septemet, xAI also signed a deal with the General Services Administration, an administrative arm of the federal government, to provide Grok to all federal agencies for forty-two cents per agency in order to “accelerate federal AI adoption.” The deal came after several AI companies — including Anthropic and OpenAI — signed deals with the General Services Administration to offer their technology to agencies at a steep discount. Originally, xAI was rumored to have been excluded from the deal due to Grok’s propensity for spouting pro-Hitler content, but the company ultimately won over officials.

The US Patent and Trademark Office — which reviews and grants patents and trademarks — has become a focus for AI companies, as it guides intellectual property policy related to AI and studies deepfakes and digital privacy issues. While efforts to curtail deepfakes and other AI copyright issues have mostly played out on the legislative level, the issue also falls within the office’s purview. One public relations executive, who represents famous clients harmed by deepfakes, explained at a 2024 roundtable that the office “is particularly well positioned to shape executive priorities both on legislation and on additional strategies.”

Hayes’s responsibilities at the US Patent and Trademark Office also include promoting and advancing the department’s own use of AI, for which the agency is actively soliciting vendors. While Hayes will technically be barred from participating in matters directly affecting xAI, he is still able to chart out the agency’s AI use while his former employer tries to win government AI contracts. (According to Hayes’s ethics waiver, as of November, the Commerce Department did not yet have any contracts with xAI.)

Hayes’s waiver is not the only dubious exemption issued to an official charged with overseeing AI adoption. David Sacks, Trump’s AI and crypto czar, received the same broad waiver as Hayes in March for his venture capital holdings, shortly after Trump fired the government’s former top ethics counsel, Hampton Dellinger. Many of the administration’s political appointees were plucked from the same businesses they are now directly overseeing.

Before Musk departed as chief of the newly created Department of Government Efficiency (DOGE), he targeted other watchdogs investigating the privacy and copyright issues presented by generative AI. The US Copyright Office was apparently targeted by Musk in May 2025 over its leadership’s resistance to AI-friendly copyright policy. In July 2024, the agency issued a report calling for Congress to pass a law to regulate deepfakes, with assistance from the US Patent and Trademark Office and the Federal Trade Commission, the nation’s top antitrust agency. (Bipartisan federal legislation cracking down on deepfakes and nonconsensual sharing of intimate images was signed into law in May 2025.)

Trump fired the acting register of copyrights and the agency’s deputy librarian at the beginning of May — a move allegedly supported by Musk and DOGE — though his replacements have historically not been as friendly to Big Tech as Musk may have hoped.

Now, though, as Grok faces mounting scrutiny over its deepfakes, Musk appears to have at least one ally in the Department of Commerce.

Generating AI Policy

Before starting his new government gig in September, Hayes spent thirty years as a “strategic C-Suite executive,” according to the conflict-of-interest waiver. Much of his career he spent in media and entertainment, working for HBO and NBCUniversal, before joining X, formerly Twitter, as an executive in August 2023, nearly a year after its takeover by Musk was finalized. There, he oversaw the launch of Grok in November 2023 and claims to have generated $425 million in new revenue for the company.

Now, Hayes holds multiple titles with the Commerce Department, serving as a senior adviser to the under secretary of commerce for intellectual property and director of the US Patent and Trademark Office. He also serves as acting chief AI officer and acting chief data officer for the patent office. He’s joined the department at a time when AI companies are pushing for more favorable intellectual property policy and expedited patent review from the office.

In 2023, former president Joe Biden issued an executive order on AI, tasking the Department of Commerce and the US Patent and Trademark Office to study and implement measures to mitigate the harms of deepfakes. The following year, the Patent and Trademark Office held roundtables on the risks of deepfakes and sought comments on how best to mitigate their harms.

In the fall of 2023, the Department of Commerce also opened the US AI Safety Institute (since renamed under Trump the Center for AI Standards and Innovation).

Under Biden, the US Patent and Trademark Office issued initial guidance on issuing patents for inventions created by or involving AI. The agency’s February 2024 draft guidance emphasized that the patent system was “designed to encourage human ingenuity” — and therefore did not permit patents “on [inventions] that are not invented by natural persons.”

This fall, under Trump nominee John Squires, a decorated patent lawyer who cofounded a financial technology company, the agency reversed course, rescinding the Biden-era approach and proposing new guidance for AI-invented patents. Under the new directive, there would be “no separate or modified standard for AI-assisted inventions.”

This summer, as Grok began to come under fire for generating racist content, Trump signed a law requiring platforms to remove harmful deepfakes when notified. The measure, to be enforced by the Federal Trade Commission, will take effect in May. But uncertainty remains about whether the legislation will successfully hold Big Tech giants like X accountable for their role in spreading harmful content.

Musk, for his part, has joked about the outrage over X’s deepfakes. On December 31, in response to a Grok-generated image of a man in a bikini, the billionaire instructed his AI chatbot to “Change this to Elon Musk.”