OpenAI Lawyers Allege Shivon Zilis Served as Musk Liaison

WIRED reports that messages and witness testimony introduced at trial in the lawsuit commonly called Musk v. Altman show Shivon Zilis acting as an intermediary between Elon Musk and OpenAI. WIRED reports that Zilis joined OpenAI as an adviser in 2016 and served on the nonprofit board as a director from 2020 until 2023. WIRED reports that trial exhibits include a February 16, 2018 text from Zilis to Musk saying, "Do you prefer I stay close and friendly to OpenAI to keep info flowing or begin to disassociate?" and a response from Musk about moving "three or four people from OpenAI to Tesla." WIRED reports Musk testified with varying descriptions of Zilis, including "chief of staff," "close adviser," and that "we live together, and she's the mother of four of my children." WIRED reports Zilis told OpenAI's attorneys she became romantic with Musk around 2016. Reporting by Yahoo Finance, citing the New York Times, documents that in 2017 Jared Birchall, an executive in Musk's orbit, registered a company described as a for-profit version of OpenAI and that records allegedly show Musk sought a 50% to 60% equity stake.

Public reporting illustrates how personal relationships and informal communications can become evidentiary anchors in corporate-governance litigation. Observers of AI governance disputes should note that board minutes and private messages often carry outsized weight when litigants dispute fiduciary purpose and transfer of control.

Comparable governance controversies in tech show that ambiguous dual roles (advisor, board member, external executive) complicate conflict-of-interest assessments and oversight. Organizations with mixed nonprofit-commercial structures face heightened scrutiny over informal channels between founders, board directors, and outside corporate affiliates.

For practitioners: Watch trial filings and exhibits for additional concrete timelines, recruitment proposals, and documented transfers of staff or IP; these items will matter more than characterization in testimony. Also follow whether independent directors, governance policies, or donor/partner agreements are cited as disputed grounds in evidentiary filings.

The reported evidence in WIRED and related coverage underscores that governance disputes in AI can hinge on interpersonal networks and documentary traces rather than high-level mission statements alone. Practitioners advising or operating hybrid nonprofit-commercial AI efforts should expect that such traces will be central if governance conflicts become litigated.