For decades, applications like Microsoft Word and email clients were synonymous with professional writing.
It was easy to assume that the amount of text people produced rose and fell with the use of those tools.
That assumption no longer reflects how work gets done today.
People haven’t stopped typing. In fact, by almost every available measure, they are producing more written content than at any other point in history.
What has changed is not the volume of writing but where it happens and, increasingly, who it is written for.
A growing share of professional writing is no longer directed at other people, but at machines.
The New Category: Writing for Machines
The shift began in a way few people anticipated.
Every interaction with an AI assistant starts with a prompt,
an instruction, question, or cue used to guide or cause a reaction.
Users describe goals, provide context, refine instructions, request revisions, evaluate results, and repeat the process, often several times for a single task.
Every one of those steps involves writing.
The scale is remarkable; OpenAI’s CEO has stated that ChatGPT now handles more than 2.5 billion prompts per day, or roughly 1.7 million every minute.
OpenAI’s own usage research reported that users were sending approximately 18 billion messages per week by mid-2025, with nearly three-quarters of conversations focused on practical guidance, information retrieval, or writing itself.
By early 2026, the service reported around 900 million weekly active users.
Competing AI assistants from Google, Anthropic, and Microsoft add hundreds of millions more interactions
Rather than eliminating writing, generative AI has created an entirely new form of human-to-machine communication.
The clearer and more precise the text, the better the outcome.
Quietly but significantly, the keyboard has become the primary interface between human intent and digital execution, and prompt writing has emerged as a professional skill in its own right.
The great migration of professional writing
This shift extends far beyond AI tools. A decade ago, the typical knowledge worker spent much of the day in Word documents and email. Today, that same person is still writing constantly, but inside collaboration platforms,
messaging apps, project management tools, customer systems, and shared documents.
Professional writing has not disappeared. It has fragmented across dozens of digital environments.
What once existed as formal documents and emails now spans chats, comments, tickets, prompts, notes, and collaborative workspaces.
The Privacy Question
All of this writing has to go somewhere, and increasingly it ends up on someone else’s servers.
Messages, documents, prompts, and notes are routinely sent to remote infrastructure for processing, synchronization, and AI-powered features.
For organizations operating in regulated industries or handling sensitive information, this raises important questions about data governance, confidentiality, compliance, and intellectual property protection.
One company addressing part of this challenge is Lightkey, an Israeli software company focused on writing assistance.
Its technology works completely offline, learns an individual user’s writing patterns, and offers inline text predictions and spelling corrections accordingly,
making repetitive writing tasks significantly faster and more efficient while helping users maintain their personal writing style.
“For decades, the ability to articulate an idea clearly was a communication skill. Today it is also becoming an operational skill.” Says Eran Brauer, co-founder of Lightkey.
“As more AI-based systems enter the workplace, the ability to write well is not just a way to convey a message to others - it is the way to make systems work for you.
Adapting to the new age of content creation, Lightkey’s upcoming edition will allow the same native prediction and forward-facing spelling correction experience in essentially every desktop and/ or web application.