‘Directionally very bad’: what Elon Musk’s trial against OpenAI can teach you about corporate communications

Two days before jury selection for the trial began, Elon Musk sent a text message to Greg Brockman regarding a possible settlement. When Brockman proposed that both sides drop their claims, Musk replied: "By the end of this week, you and Sam will be the most hated men in America."

The message, revealed in a filing, framed the approach OpenAI's attorneys would take throughout the trial: Musk's texts revealed outsized personal grievances against Altman, and his suit was a case of "sour grapes" for missing out on OpenAI's later success. Exercising vigilance over tone in private channels could spare trouble down the line, should conversations arrive before a jury. 

Read more: ChatGPT Atlas vs Google Chrome: the battle for your browser begins

Sam Altman's frantic messages to then-interim CEO of OpenAI, Mira Murati, shortly before he was ousted from OpenAI's board, became the most-memed exchange of the entire trial. "Can I come in?" Altman wrote, asking to enter OpenAI's offices; Murati's reply, "They don't want you to," was screenshotted, captioned, and reposted across X within hours of the exhibit's release to the public. One key line from Murati, that the decision was "directionally very bad" for Altman, quickly entered the lexicon of social media users.

As the fast-moving governance crisis unfolded, Altman's texts turned a delicate, if slightly embarrassing moment into a viral artifact. Personnel decisions, regulatory concerns, tensions, and competitive assessments are all high-stakes situations that can be easy to misinterpret in text, and almost always better discussed verbally. A five-minute chat can accomplish what a carefully worded text cannot: it lets executives discuss without their initial reactions becoming evidence.

Most workplace communication is conducted through digital platforms, including text messaging applications, email, and internal messaging systems like Slack. Companies typically maintain data retention policies governing the storage of such records, but during litigation, individuals may be required to produce relevant communications that have been preserved. In Elon Musk's case, messages dating back to 2014 were introduced as evidence and discussed in open court.

All work-related communications, regardless of platform, can be used in litigation, and mixing personal banter with business decisions in the same thread means it all becomes part of the corporate record. The cleanest approach, depending on where you work, could be as simple as having a dedicated device or account for work, and allocating specific apps for work communications.

Read more: What to know about Scale AI, the data company powering OpenAI, Meta and Google

In an April 2022 email to Microsoft executives, Nadella wrote of OpenAI's other backers: "I just don't get their incentives. I would rather have OpenAI employees get money at appropriate delivery milestones… vs these VCs."

Shared with a small internal circle of fellow staff, the line reflected Nadella's strategy; surfaced in court, it became a public characterisation of named investors. Written assessments of partners, counterparties, and competitors should be drafted on the assumption that the people being assessed will eventually be among the readers.

Read more: What is an ‘infinite workday’? This latest Microsoft study shows the rise of after-hours work, constant communication and weekend catch-ups
 

From the moment he filed suit in 2024, Musk's public persona was consistent: he was a whistleblower and early backer of OpenAI, defending its original nonprofit mission against a leadership that had abandoned it for private gain. He testified that he "was a fool" who unwittingly donated US$38 million to create a corporate behemoth.

His emails told a different story. In 2017 and 2018, Musk had privately proposed folding OpenAI into Tesla, thereby taking a majority stake and the title of CEO. When the board declined, he left. Yet he remained in contact with Altman, and in 2020 posted on X: "This does seem like the opposite of open. OpenAI is essentially captured by Microsoft."

Executives now post on X, Threads, and LinkedIn while texting privately about the same subjects. Inconsistency between platforms may become obvious when placed side by side, and when exposed, can leave you vulnerable to criticism.

Between Musk and the many defendants of the case, credibility went on trial. Musk's lead attorney spent long stretches of cross-examination pressing Altman on whether he was "completely trustworthy" and whether he had "misled people with whom you do business". Altman conceded he had been "repeatedly called deceptive and a liar," referring to OpenAI co-founder Ilya Sutskever's 52-page memo accusing him of "a consistent pattern of lying, undermining his execs, and pitting his execs against one another."

Conversely, Altman's team cast Musk as a rival seeking to reclaim control of a team he had previously walked away from while establishing a competitor, now known as xAI. Through the process, they also argued that Musk's use of disappearing messages on Signal suggested an intent to destroy evidence. 

Executives may not be able to choose how they're characterised in a courtroom or a news cycle. They can choose, over time, whether their trustworthiness is backed by a paper trail that favors their version of events.

Read more: Meet Nguyen Hong Minh, education expert on a mission to take Vietnamese students to Microsoft’s global AI stage
 

Musk's landmark court case failed to create any legal precedent—after all, the case ended on a procedural finding, not a ruling on its merits—but it successfully showed a generation of tech employees that the way they communicate, warts and all, can come back to haunt them in a courtroom. Musk and Altman might have learned this lesson the hard way. 

READ NOW

OpenAI founder Sam Altman joins The Giving Pledge, committing to donate at least US$1 billion to charitable causes

41 per cent of employers plan to cut roles that AI can replace by 2030

ChatAndBuild founder Christel Buchanan on how AI is changing who builds the next big app