Workplace Texts Can Become Evidence. The OpenAI Trial Shows Why

The OpenAI trial has turned a private workplace exchange into a public lesson: what feels like a temporary text can become a lasting record.

A 2023 text exchange between Mira Murati and Sam Altman surfaced during the Musk v. Altman trial, drawing attention because it took place during one of the most turbulent moments in OpenAI’s recent history. Altman had been pushed out as CEO, Murati was serving as interim CEO, and the company’s board was still weighing what came next.

The exchange, as described in the source report, showed Altman asking for a meeting with the board and Murati relaying that the board did not want him back at that moment. The details are specific to OpenAI, but the broader workplace lesson is not: sensitive work conversations that happen in writing can later be reviewed, produced, quoted, interpreted, or used against someone in a very different setting.

That does not mean every tense message is destined for a courtroom. Most workers will never experience anything close to the OpenAI leadership fight. But the episode is a sharp reminder that modern work leaves a thick paper trail, even when no paper is involved.

Why Workplace Texts Feel Private But Often Are Not

Text messages can feel different from Slack, Teams, email, or a shared document. They usually live on a phone. They may be mixed with family messages, personal plans, photos, and everyday private life. That makes them feel informal and separate from work systems.

Legally and practically, that separation can be thinner than many employees assume.

If a message is about work, it may matter less where it was sent and more what it was about. A text sent from a personal phone, a WhatsApp message, an Instagram DM, or a note created by an AI meeting assistant may all become relevant if they touch the subject of a workplace dispute, investigation, lawsuit, or regulatory review.

An employment attorney quoted in the source report described the key issue as relevance. In litigation, a request may seek information that relates to the case regardless of whether it sits on a company laptop, a personal phone, or a consumer messaging app. That does not mean every private message is automatically collected. It does mean that using a personal device for work does not guarantee that work-related messages stay out of reach.

The risk is not only legal. Many companies reserve the right to monitor activity on work systems, retain workplace communications, and review content during internal investigations. Employees may understand that emails are archived, but they may forget that chat tools, shared drives, calendar entries, meeting transcripts, and project-management comments can also create records.

That gap between how casual a message feels and how permanent it may become is where many problems start.

The Personal Phone Problem

The cleanest boundary is also the least convenient one: keep work and personal communications separate.

The attorney quoted in the source report recommended using separate devices when possible, with work content confined to a work phone. That advice may not be realistic for every employee. Some companies do not issue phones. Some workers are expected to use personal devices for two-factor authentication, scheduling, calls, or quick replies. Others prefer one phone because carrying two is annoying.

Still, the principle is useful even when a second phone is not practical. The more work content you place on a personal device, the more complicated that device becomes if a dispute arises.

A personal phone may contain years of unrelated private material. If it also contains relevant work texts, DMs, files, or app data, separating the two can become difficult. That is why workers in sensitive roles should think carefully before letting work conversations spill into personal channels.

For employees, that means asking a few concrete questions:

• Am I discussing company decisions, personnel issues, legal concerns, customer data, strategy, or confidential information on a personal device?
• Would this message create problems if it appeared in an investigation file or court exhibit?
• Is there a company-approved system where this conversation should happen instead?
• Does my employer have a policy about texting, personal devices, retention, or messaging apps?
• Would a phone call or in-person conversation be more appropriate for this topic?

The point is not to avoid accountability. Important workplace decisions often need a record. The point is to match the communication channel to the subject. A routine scheduling note is different from a discussion about layoffs, executive conflict, discrimination complaints, trade secrets, board deliberations, or legal exposure.

AI Tools Add Another Layer of Records

Hybrid work has made meetings easier to record, transcribe, summarize, and search. AI notetakers and meeting assistants can be useful, but they can also create records that employees may not think about in the moment.

The source report notes that AI notetakers are increasingly present in work calls and may create written or audio records. That point should be treated carefully because the details depend on the specific tool, settings, company policy, consent rules, and retention practices. Still, the practical concern is real: tools that summarize meetings may preserve sensitive discussions in places employees rarely check.

A meeting that once would have disappeared into memory may now generate a transcript, a summary, action items, searchable keywords, or a recording. Those materials may be stored by the employer, the tool provider, or a connected workspace. They may also be shared automatically with attendees or added to a project channel.

That can be helpful for ordinary project work. It can be risky for sensitive conversations.

Before discussing confidential or legally sensitive topics in a meeting with AI tools present, workers and managers should know:

• whether the call is being recorded or transcribed;
• who receives the transcript or summary;
• where the output is stored;
• how long it is retained;
• whether the tool uses meeting content to improve its services;
• who has admin access to the workspace where the record is stored.

For companies, this is a policy issue, not just an employee habit issue. If AI tools are allowed in meetings, employees need clear rules about when they can be used, when they must be disabled, and what happens to the records they create.

When Writing Helps And When It Hurts

There is a trap in any conversation about digital hygiene: it can sound like the advice is to never write anything down. That is not realistic, and it is not always wise.

Written communication is essential for many parts of work. It clarifies instructions, documents decisions, preserves institutional memory, and protects people when expectations are disputed. A good written record can prevent confusion and reduce risk.

The problem is careless writing, not writing itself.

A rushed text can compress context, emotion, and ambiguity into a few lines. A sarcastic remark may read differently months later. A vague comment may look more suspicious than it was. A message sent in frustration may outlive the frustration and become the version of events other people analyze.

The safer habit is to write as if the message may one day be read by someone outside the conversation. That does not mean sounding robotic or defensive. It means avoiding speculation, personal attacks, exaggeration, and half-formed thoughts on sensitive subjects.

For example, there is a difference between these two approaches:

• Riskier: sending an emotional side-channel message about a disputed personnel decision.
• Safer: documenting the business reason for the decision in the proper system, using direct and factual language.

The same principle applies to legal issues, internal complaints, investigations, conflicts of interest, executive decisions, customer problems, and anything involving confidential company information.

Practical Digital Hygiene For Employees

Digital hygiene at work is not about paranoia. It is about reducing avoidable exposure.

The first rule is to know where your work conversations are happening. Many employees now operate across email, Slack, Teams, text messages, shared docs, voice notes, calendar comments, ticketing systems, AI summaries, and private messaging apps. That spread makes it easy to lose track of what has been said and where it lives.

The second rule is to keep sensitive subjects out of casual channels. If a topic involves legal risk, personnel decisions, confidential company information, board matters, security incidents, or a serious complaint, pause before texting.

The third rule is to respect retention rules. The source report says the attorney it quoted recommends periodically deleting personal communications such as text messages. That advice should be handled cautiously. Employees should not delete messages to hide evidence, evade an investigation, violate company policy, or ignore a legal hold. Once litigation, an investigation, or a preservation obligation is reasonably in view, deletion can create serious problems.

A more careful version of the habit is this: maintain normal privacy and cleanup practices before a dispute exists, while following company policy and legal obligations when they apply.

Workers can also reduce risk by separating accounts and apps. Use company-approved tools for company business when possible. Avoid forwarding work material to personal email. Avoid downloading sensitive files to unmanaged devices. Be cautious about screenshots, copied text, and pasted prompts in AI tools.

That last point matters. If an employee pastes confidential work material into an AI chatbot or summary tool, the privacy and retention rules depend on the tool, account type, settings, and employer policy. Workers should not assume that a prompt disappears just because the interaction feels conversational.

What Managers And Companies Should Take From This

Employees are not the only ones responsible for cleaner communication. Companies create risk when expectations are vague.

If leaders want employees to avoid sensitive work texts, they need to provide usable alternatives. If they want employees to use approved messaging tools, those tools should be practical for the way the company actually operates. If AI meeting assistants are allowed, there should be clear rules for consent, storage, access, and retention.

A company policy that sits unread in a handbook will not change behavior by itself. The useful version is specific:

• which tools are approved for work communication;
• what information cannot be sent through text or personal apps;
• when employees must disable recording or AI transcription;
• how legal holds and investigations affect deletion;
• who to contact before discussing sensitive legal, HR, or board matters in writing.

Managers also need training because their messages carry extra weight. A casual executive text may later be treated as evidence of company intent. A manager’s offhand comment about an employee, customer, competitor, or legal issue can become more significant than intended.

The OpenAI example stands out because the people involved are well known and the stakes were unusually high. But the pattern is ordinary. Work conflict happens. People text quickly. Messages survive.

The Core Lesson

The lesson from the OpenAI trial is not that employees should stop communicating. It is that workers should be more deliberate about where and how they communicate when the subject is sensitive.

A personal device is not a magic shield. A private app is not always private in a legal dispute. AI tools can create records workers may forget exist. Deleting messages is not a universal fix and can be risky when preservation duties apply.

The safest everyday standard is simple: if you would be uncomfortable seeing a message quoted in a workplace investigation, court filing, or board packet, slow down before sending it.

Some conversations belong in writing because a clear record is necessary. Some belong in a formal system because the company needs accountability. Some are better handled by phone, in person, or with legal or HR guidance before anyone starts typing.

The hard part is recognizing the difference before the message is sent.