When AI Writes, Humans Must Think

A recent article in The Economist pointed to a striking development in the AI boom: technology companies are hiring philosophers. The story was quickly picked up by others, including The Atlantic and The New York Times. At first glance, it sounds like a charming paradox: the most advanced technology companies in the world turning to one of the oldest disciplines in the world.

But the paradox disappears when technology starts making decisions. At that point, the hardest questions are no longer only about capability. They are about consequence.

Philosophical Questions

The hardest questions in AI are no longer only technical. They are philosophical. What is true? What is fair? What should a system optimize for? When should it refuse? Whose values are being encoded? What happens when a technically correct answer is socially wrong?

These questions may sound abstract until systems begin to influence real choices, real relationships and real reputations. Then they become operational. That is why communications leaders should pay attention. The same shift is now reaching our function.

For years, communications has been organized around output: press releases, speeches, posts, Q&As, briefings, campaigns, reports. AI will not make all of that disappear. But it will make much of it cheap, fast and increasingly generic. Drafting, summarizing, translating, adapting, monitoring and reporting will move to agents. The content machine will become largely synthetic.

That creates an uncomfortable question. If AI can produce the output, what is the communicator for?

At Scriptorium, we see this tension in our experiments with synthetic communications teams. The agents can already generate options, draft responses, map stakeholders and simulate reactions. But the quality jump does not come when the agent writes more. It comes when the human thinks better. And that thinking is increasingly philosophical.

Question of Truth

First, there is the question of truth. Philosophers call this epistemology: how do we know what we know? For communicators, this is becoming a daily discipline. Which signals are real, and which are noise? What can we infer from a stakeholder reaction? What is the difference between a pattern and a conclusion? When does a dashboard create insight, and when does it create false confidence?

Second, there is the question of ethics. What is the responsible thing to do? Not only what is legally defensible or reputationally convenient, but what the organization can credibly stand behind. Should we speak now, or wait? Should we personalize this message, or would that feel manipulative? Should an agent be allowed to simulate stakeholder reactions, and where should the human boundary remain?

Third, there is the question of taste. This may be the most underestimated one. As machines generate more language, the premium shifts to human discernment. Does this sound true? Does it carry the right weight? Does it feel like leadership, or like corporate wallpaper? Is the silence in the message as important as the sentence itself?

This is where humans still matter. Not because we are better at producing words, but because we bring experience, judgment and taste. We understand context. We sense when something is technically right but reputationally wrong. We know that speed can destroy trust, and that a message can be polished and still feel false.

In Scriptorium's work on the future structure of the communications department, the applied philosopher is no longer a poetic idea. It is becoming a practical role: someone who helps the organization think through truth, legitimacy, values, stakeholder impact, human voice and the boundaries of automation.

This may sound abstract. It is not. It may become one of the most practical roles in the future communications function. Because if agents take over production, communicators will have to prove they are more than producers.

Speed Without Judgment

The future communications function cannot stop at the agentic newsroom. That model may create speed, scale and consistency. But speed without judgment is not strategy. Scale without legitimacy is not trust. And consistency without taste is just corporate wallpaper.

The real opportunity is to build a reputation intelligence and trust function. A function where agents provide sensing, drafting, simulation and measurement, while humans provide interpretation, ethics, relationships, voice and accountability.

This is exactly the shift the Scriptorium Initiative sees in the future communications function: more automated in execution, but more human in purpose.

The real lesson from Silicon Valley is not that every organization should put a philosopher on the payroll. It is that AI will expose the difference between people who create output and people who create meaning. For communication leaders, this is the uncomfortable question. If agents can draft, summarize, simulate and optimize, what remains distinctively human in communications?

In a world where machines can generate almost any message, the human premium will sit with those who know what is true, what is wise, what should be said, what should not be said, and what the organization must be willing to stand behind.

Nanne H.C. Bos

Chief Communications & Brand Officer | Corporate Affairs, Reputation & Transformation | AI, Culture & Leadership

Follow Nanne Bos on LinkedInVisit Scriptorium Initiative