A personal anecdote, three management lessons, an inspiring closing line: on LinkedIn, leaders have their formulas for communicating. Artificial intelligence can now reproduce them in seconds. When anyone can publish a flawless text, why believe one rather than another? For business leaders and their communications teams, the challenge is no longer simply to speak out. It is to show that a real person, accountable for their choices, stands behind the words.
By September 2026, this question could become central to professional communications. This forward-looking analysis draws on developments already documented, notably the arrival of generative writing tools on LinkedIn in 2023. It assumes neither a measured, widespread adoption of automated posts nor an algorithm change that would still need to be verified.
A writing assistant for everyone
The shift is not being driven solely by external tools. In 2023, LinkedIn introduced AI features designed, among other things, to improve profiles and job postings, followed by assistance with drafting posts for some Premium subscribers. The platform also launched collaborative articles, started with AI and then expanded through human contributions. Text generation has thus become part of how the network is used.
For a communications department, the benefits are tangible. Working from an interview with a leader, a tool can suggest an outline, shorten a technical explanation or produce several opening lines. Someone who struggles with writing can express their ideas more clearly. A small business can achieve a standard of presentation once reserved for organisations with an editorial team.
But this reduction in production costs creates a temptation: publishing because it has become easy. A factory visit becomes a lesson in leadership; a meeting, a case for collective intelligence. By turning every moment into content, communications can lose the very thing that justified them: information, a conviction or an experience genuinely useful to the reader.
The problem is not just who does the writing
Leaders have never communicated entirely on their own. Speeches prepared by colleagues, reworked opinion pieces, edited interviews: assisted writing long predates generative models. Requiring every post to be typed personally by the person whose name appears on it would therefore be a false test of authenticity. The decisive question is ownership: does the leader understand, approve of and stand ready to defend what is published under their name?
AI does, however, change the scale of the problem. It can quickly manufacture a plausible voice, complete with calculated hesitations, confidences and enthusiasm. A text can suggest intimate personal experience without being based on anything actually lived. The risk then goes beyond bland writing: it affects the implicit contract with the reader, who interprets a first-person account as testimony.
Even so, a few turns of phrase cannot reliably establish where a text came from. A repetitive structure or smooth vocabulary proves nothing. Blanket suspicion would lead nowhere. It is better to assess observable elements: the specificity of the account, its consistency with actions and the author’s ability to respond to objections.
Three ways to rebuild credibility
Tell stories a model cannot supply
An experience is worth more than a ready-to-publish moral. Take a fictional example: a business leader explains why her company postponed a service launch after a disappointing customer test. What matters is not a line about the courage to fail, but the signals that were overlooked, the trade-off that was made and the decision that followed.
Useful details are not necessarily trade secrets. They might involve a scheduling constraint, an abandoned assumption or a disagreement between departments. AI can help organise this material; it must not invent it. A good editorial habit is to ask, before drafting: what have we learned that we did not know before?
Own a position and its limits
It costs little to advocate innovation, listening or trust. It takes greater commitment to explain the circumstances in which you refuse to automate a service, why you retain a less profitable operation or what a return-to-office policy is really trying to solve. A credible position sets out its compromises, not just its benefits.
This does not mean courting controversy. Escalating opinion can become another form of mass production. A well-grounded statement distinguishes between facts, personal judgement and uncertainty. It also makes room for acknowledging change: admitting that a decision did not have the intended effect can sometimes build more trust than a succession of perfectly narrated successes.
Stay involved after publication
Comments often demand what a post allows its author to avoid: clarification, justification and listening. An employee points out a gap between the message and their daily experience; a customer asks for an example; a professional challenges a conclusion. How these contributions are handled provides a tangible indication of the named author’s engagement.
A team can sort questions and prepare talking points. But delegating the entire exchange while simulating a personal conversation creates a risk of deception. On sensitive issues, a slower response that has genuinely been approved is better than artificial availability. And not every criticism calls for a defence: some deserve a visible correction.
Move from a production line to an editorial process
The organisational response is less about hunting for suspicious words than establishing simple rules. The work begins by gathering facts from the leader and the relevant teams. It continues with checks on anecdotes, data and any permissions required. Entering information about a customer or employee into a tool also raises confidentiality concerns.
- Ground: start with a verifiable fact, a decision or a question actually encountered.
- Set boundaries: allow help with wording, never the invention of experiences or testimony.
- Approve: obtain the named author’s informed consent, not a distracted sign-off.
- Follow up: organise responses and correct errors after publication.
Transparency should be proportionate to the use made of AI. Correcting syntax does not raise the same issues as generating most of a personal account. A sound rule is never to imply that an experience or interaction took place when it did not. Performance, too, deserves to be reassessed: a substantive conversation may matter more than broad exposure that leads nowhere.
What next? If generated posts continue to proliferate, scarcity could shift from text to evidence of involvement. Leaders will not necessarily have to write alone or give up AI. Above all, they will need to remain present in their own communications: supplying the material, taking responsibility for trade-offs and accepting challenges. On LinkedIn, credibility may belong less to those who publish best than to those who have something to share — and stay to discuss it.
