From Witness to Author

I. Thirty Years in the Room

I started in a newsroom in Georgetown, Guyana, before I owned any part of one. I learned the trade the old way: chase the story, check it twice, run it only when you can stand behind it. From there I founded newspapers and went on to originate new media firms as a media entrepreneur, in Guyana and in Canada. I pioneered documentary style television in Guyana when almost no one there was doing it. Three decades in, I have watched this industry reshape itself three or four times, and the numbers now confirm what those of us inside it have felt for years.

Newspapers in the United States are closing at a rate of more than two a week. Northwestern University's Medill school, which has tracked the industry for two decades, counted 136 closures in the past year alone, part of nearly 3,500 newspapers and nearly 270,000 newspaper jobs lost since the early 2000s. Almost 40 percent of local newspapers in the country have vanished since 2005. At the same time, an Amazon Web Services study cited in the Reuters Institute's 2026 industry forecast found that 57 percent of content published online is now written or translated by a machine. One curve is falling. The other is rising. Media sits exactly where they cross.

II. The Council Meeting

Picture a city council meeting on a Tuesday night. A reporter used to be sent to sit through three hours of zoning arguments, take notes by hand, miss the side conversation in the hallway, get one good quote, and call it coverage. That reporter was a witness, but a limited one. One set of eyes, one set of ears, one attention span fading by hour two.

This is no longer theoretical. Chalkbeat, a nonprofit education newsroom, now uses an AI transcription tool called LocalLens to search and index school board and town hall meetings across the districts it covers. One of its reporters has used the archive to locate a student source's testimony at a meeting she had never personally attended, work she told Nieman Lab would otherwise have taken an entire day per meeting just to review. A separate study of the Grand Rapids Documenters program found that AI assisted meeting notes scored higher for accessibility, readability, and inclusion of public comment than the city's own official minutes. Researchers building a system called PUBLICSPEAK have already pulled public remarks out of more than 30,000 minutes of government meeting footage across seven U.S. cities, the kind of coverage no single newsroom could staff by hand.

For an event like this, plain, procedural, fact heavy, AI is not a weaker witness than a person. The evidence says it may be the stronger one.

I want to be honest about that, because I was not honest about it the first time I wrote on this subject. I treated witnessing as something only a human body could do, and I was protecting an old idea of the journalist that the data was already starting to outgrow.

III. Five Questions, One That Stays Ours

Every reporter learns the same five questions before anything else. What happened. Where. When. Who was involved. How did it unfold. Get those right and you have a story that holds up.

AI is fast becoming excellent at the first four questions and most of the fifth. This is not a small claim. The Reuters Institute's 2026 survey of newsroom leaders found that fewer than four in ten senior news executives feel confident about journalism's prospects this year, a confidence gap tied directly to AI driven answer engines, which publishers now expect to cut referral traffic to their own websites by 40 percent over the next three years. The fact layer, what, where, when, who, and most of how, is being absorbed faster than most newsrooms can adjust to.

One question stands apart from the rest, because it was never really a reporting question to begin with. Why does this matter to the family two streets over. Why should anyone act, vote, write, or change a plan because of what happened in that room.

Why is a judgment, not a record. It needs a person who has something at stake, a stand to take, a community to answer to. AI can hand you the facts of the meeting beautifully. It cannot tell you why your life should bend around them.

IV. From Witness to Author

For most of my career, media's job stopped at the first four questions plus how. Get the facts, write them clean, run them before the competition got there. The witness and the reporter were one job, and a good reporter was someone you could trust to do it honestly.

AI now does a version of that job, often a faster and more complete version. That should not threaten the people who used to do the witnessing. It should promote them. Once the facts of the council meeting are gathered, checked, and laid out, a person is free to do the work AI cannot touch: deciding what the facts mean, building an argument from them, placing them inside a larger story about who we are and where we are headed.

Alba's point is the same one I am making here, from inside a different career. Journalism's worth was never really about how much it could produce. It was about the rigor standing behind whatever it produced, the willingness of a person to look at the world directly and put their name on what they saw. AI changes the arithmetic of how fast the facts arrive. It does not change who is willing to stand behind the meaning of them.

I call this the move from witness to author. The witness reports what happened. The author decides what it means and what to do next. Journalism spent a century and a half mostly stuck at the witness stage because gathering the facts took almost everything a newsroom had to give. AI changes that math. The facts get gathered faster, which means more of a person's time and judgment can go into the part only a person can do.

V. From Audience to Community Designer

This same shift runs all the way down to the individual, and the local news numbers explain why it has to. With 213 U.S. counties now without any local news source, 1,524 more counties down to a single outlet, and roughly 50 million Americans living with limited or no access to local journalism, the old assumption, that someone else is always in the room watching on your behalf, no longer holds for a growing share of the country. For most of media history, an ordinary person received the report and that was the end of their part in it. Read the paper, watch the broadcast, form an opinion privately, move on.

AI removes the bottleneck that kept people sitting in the audience seat. The facts of the world, the what, where, when, who, and how, now arrive gathered and organized faster than any one person could gather them alone. What is left for that person to build is the part that was always theirs and rarely had the time for: a response. Their own account of what the news means for their life, their family, their work, their community, written and kept and shared in their own words.

That is the premise behind QwaiAI. Not a tool that writes the response for you, a guide that helps you build it yourself, so the Smart Book you keep is not a diary of what happened to you but a record of what you decided it meant and what you did about it. Media's job is no longer only to report the world to an audience. It is to help that audience become authors of their own answer to the world, and then to help authors who share an answer find each other and build something together.

VI. Three Duties

Media in this age still has three duties, though I would state them differently now than I once did, and the people doing this work at the highest level are already living out all three.

First, deploy AI as a witness, fully and without sentiment. Use it to cover the rooms a newsroom could never staff, to cross check the record, to catch the contradiction a tired reporter would miss. Resisting this on principle protects nothing except a slower, thinner account of the world.

Second, keep the why as human territory, and defend it. Taneth Evans, Head of Digital at the Wall Street Journal, told the Reuters Institute that newsrooms are now waking up to the need to double down on "the things that make us valuable and unique." Every report AI produces still needs a person willing to say what it means and to put a name to that judgment.

Third, build the tools that turn a fact into a response, and a response into a community. Sophia Phan, Senior Audience Editor at Nine in Australia, put the task to the Reuters Institute plainly: as search traffic fades, newsrooms have to find ways to "build experiences that drive habitual use" rather than one off visits. A report that ends at the facts leaves people exactly where they started, informed and inert. Media's work now is to hand people something to build with: a way to write their own account of what they read, connect it to others doing the same, and turn private understanding into shared direction.

VII. Where This Leaves Us

I founded my first newspapers because I believed people deserved an honest account of the world around them. I still believe that, and I no longer think the account is the hard part. AI can give us the account faster and more completely than any newsroom I have ever run. The data backs that claim up, even when it is uncomfortable to read.

What AI cannot give us is the why, and it cannot give us each other. Judgment and community were always the real work underneath the reporting, even when newsroom economics buried that work under the daily grind of gathering facts. The mission underneath everything I have built, Qualped, QwaiAI, the novel, the magazines, has always been the same one: help a person move from receiving the world's report to authoring their place in it. AI did not end that mission. It cleared the desk so the real work could finally begin, for journalists and for everyone else trying to make sense of a Tuesday night council meeting that will, in some quiet way, change their street.