BoSacks Speaks Out: A Manifesto for an Industry That Must Wake Up

BoSacks Speaks Out: A Manifesto for an Industry That Must Wake Up

I have spent more than five decades watching publishing absorb one technological upheaval after another. Phototypesetting, desktop publishing, digital production, the internet, email, mobile devices, and social media each changed how we produced and distributed our work.

AI feels different because it has moved beyond the machinery of publishing and entered the editorial room. It does not merely help us manufacture or deliver the words. It can now help research, organize, edit, illustrate, summarize, and write them.

My relationship with AI is not theoretical. I have spent more than fifty years making a living with words while wrestling with dyslexia, unreliable spelling, and no great friendship with formal grammar. Finding the thought was seldom my problem. Presenting it cleanly was another matter.

Today, AI helps me cross that gap. I bring the experience, memories, opinions, questions, and occasional outrage. AI helps me organize the argument, identify repetition, challenge weak reasoning, and correct the spelling.

I do not ask it to decide what I believe. I use it to help me express what I believe. The ideas remain mine, as does the responsibility for every word I publish.

That daily experience has allowed me to see both AI’s remarkable abilities and its very real failures. It can clarify a confused argument in seconds, then confidently invent a fact that never existed. It is a powerful assistant, an unreliable witness, and a terrible choice for unsupervised management.

I repeat this personal history because it explains my perspective. I am neither cheering from the sidelines nor throwing rocks at the machinery. I am using the technology, learning from it, questioning it, and trying to understand what it means for the industry in which I have spent most of my life.

That experience has persuaded me that publishing must abandon several comforting assumptions about what AI can and cannot do. We need less reassurance, more honesty, and a serious plan for the industry being built around us.

We keep telling ourselves that human writing possesses some mysterious essence a machine cannot touch. We insist that readers can always detect the difference. We repeat the reassuring claim that clarity, originality, and insight remain uniquely human qualities.

I wish that were true.

AI is already producing clear, thoughtful, and occasionally insightful work. It does not do so consistently. It makes mistakes, manufactures facts, imitates voices, and can produce prose as lifeless as the instructions accompanying a new refrigerator. But it is improving, and it is improving quickly.

The gap between average human writing and machine-generated writing is narrowing. In some forms of routine writing, it has already disappeared.

That does not mean humans have become unnecessary. It means being human is no longer, by itself, a sufficient competitive advantage.

The publishing industry needs to stop reassuring itself and begin preparing itself.

So here is my manifesto.

1. Editors must stop thinking their primary job is fixing sentences.

AI can correct spelling, repair grammar, tighten sentences, suggest headlines, restructure paragraphs, and summarize long documents. The mechanical portion of editing has already been automated to a remarkable degree.

That does not make editors obsolete. It makes good editors more important.

An editor’s real value is not knowing where to put a comma. It is knowing when an argument is weak, when a source is unreliable, when a writer is avoiding the obvious question, and when a technically accurate story is nevertheless misleading.

The editor of the future will be a strategist, curator, verifier, coach, and guardian of meaning. Judgment will matter more than mechanical skill.

If your only value is fixing grammar, you are vulnerable.

If your value is judgment, curiosity, context, and taste, you may be indispensable.

2. Publishers must stop pretending that volume is a business model.

AI can produce almost unlimited quantities of content. It can generate articles, newsletters, summaries, social posts, product descriptions, and search-friendly filler around the clock.

The world does not need more content. We are already drowning in it.

What readers need are better filters. They need publications willing to decide what matters, what is true, what deserves attention, and what can safely be ignored.

The publisher of the future cannot merely be a content factory. A publisher must be a filter, a guide, and, on the best days, a lighthouse.

If your business depends primarily on producing more material than everyone else, AI will make that advantage disappear.

If your business depends on trust, usefulness, and a clearly understood relationship with an audience, you still have something worth defending.

3. Writers must stop believing that being human is enough.

It is not.

Competent, coherent, good-enough writing is no longer exclusively human territory. AI can already produce an acceptable first draft on thousands of subjects, often faster than a human writer can make coffee.

The threatened territory is not great writing. It is predictable writing.

If you write what anyone could write, using information available to everyone and offering conclusions everyone expects, a machine can probably do it faster and cheaper.

The most defensible human territory is perspective earned through experience. It is the ability to say, “I was there. This is what I saw. This is what I misunderstood then, and this is what I understand now.”

AI can imitate a voice. It cannot live a life.

It has never run a press, missed a deadline, fought with a printer, lost an advertiser, launched a magazine, buried a friend, raised a family, made a terrible decision, or learned something important too late.

Human experience still matters, but only if writers bring that experience to the page.

4. We must stop romanticizing authenticity while still protecting it.

Authenticity is an attractive word, but it is not automatically a business model. A human being is perfectly capable of producing dishonest, derivative, boring, and worthless prose.

A human byline does not rescue a bad article.

At the same time, authorship matters when readers are being asked to trust a publication, accept an argument, or respond to someone’s claimed experience. Readers deserve to know whether they are hearing from a person, a machine, or some combination of the two.

The answer is not to pretend that all human work is authentic and all AI work is fraudulent. The answer is honesty.

Tell readers how the work was produced when that knowledge is relevant. Do not present machine-generated experience as lived experience. Do not manufacture expertise. Do not create imaginary witnesses, sources, or emotions.

Authenticity is not merely a feeling. It is a relationship built on truthfulness.

5. We must redefine what human writing is for.

Human writing may become the premium tier of publishing.

It will be the work carrying the weight of memory, reputation, responsibility, and consequence. It will come from people who have something at stake and are willing to stand behind their words.

That does not mean every human writer must produce literature. Most publishing serves practical purposes, and there is nothing wrong with practical writing. But human writers must contribute something beyond competent sentence construction.

They must bring curiosity, skepticism, humor, empathy, history, contradiction, and the unpredictable connections that emerge from having lived in the world.

AI can manufacture an answer.

A writer should provide a reason to care.

6. The industry must stop defending the past and begin designing the future.

AI is not simply a threat to writers. It is a threat to every publishing process built around repetitive labor, predictable formats, and interchangeable content.

That disruption will be painful. Jobs will change. Some will disappear. New ones will emerge, although probably not on the same timetable or in the same places. Anyone who tells you the transition will be painless is selling something.

But AI also offers publishers extraordinary tools. It can help small operations perform research, analyze data, organize archives, improve accessibility, translate material, and undertake projects that were previously too expensive or time-consuming.

The question is not whether publishers should use AI. They already are.

The questions are how they will use it, what standards they will establish, and whether they will allow short-term savings to destroy long-term trust.

7. Publishing’s value must shift from production to judgment.

Routine creation is becoming cheap. Content is becoming infinite.

Judgment remains scarce.

The publisher of the future will not earn loyalty merely by producing information. The publisher will earn it by selecting, verifying, contextualizing, and explaining information.

This requires people who know their subjects and understand their audiences. It requires editors with the courage to reject plausible nonsense. It requires publishers who recognize that a million inexpensive articles may be worth less than one indispensable article.

The publisher of the future must become a maker of meaning.

That sounds like something my friend Professor Samir Husni might say, which is one reason I suspect it is true.

8. Copyright cannot become collateral damage.

The publishing industry cannot embrace AI while ignoring how many of these systems were built.

Writers, photographers, illustrators, publishers, and other creators deserve to know when their work is being used. They deserve enforceable rights, reasonable licensing systems, and compensation when their intellectual property creates commercial value for someone else.

Innovation is not a license to take whatever can be scraped from the internet.

If publishers surrender copyright because enforcing it is difficult, they may eventually discover that they have surrendered the economic foundation of their own businesses.

We can welcome useful technology without agreeing that everything ever published has become free raw material.

9. Trust must remain the business model.

Trust is not a fashionable new publishing strategy. It was always the business model.

A publication survives because an audience believes that someone responsible stands behind it. That does not mean the publication is always correct. It means errors are corrected, methods are explained, sources are examined, conflicts are disclosed, and someone accepts accountability.

AI cannot accept responsibility. It cannot be embarrassed by an error, apologize to a source, defend a decision in court, or lose sleep because it published something harmful.

People remain responsible for what they publish, even when a machine helped create it.

Every publisher needs clear rules governing AI-generated material, verification, attribution, copyright, disclosure, and human review. Those rules should be understandable to employees, contributors, advertisers, and readers.

“Everyone else is doing it” is not an editorial standard.

10. We must stop confusing effort with value.

Publishers have historically attached value to the difficulty of producing something. We employed writers, editors, designers, researchers, photographers, production specialists, and printers. The process required time, skill, money, and no small amount of caffeine.

AI changes the cost of production, but effort was never what readers bought.

Readers bought usefulness, pleasure, surprise, knowledge, identity, and trust.

The fact that something took three weeks to produce does not make it valuable. The fact that AI helped create it does not automatically make it worthless.

The only durable question is whether the result deserves the reader’s time.

That standard is simple to state and brutally difficult to meet.

The Human Fingerprint Is Fading

The human fingerprint on routine content is becoming harder to see. Publishers who pretend otherwise are not protecting writers. They are postponing an unavoidable reckoning.

AI will dominate much of the middle. It will produce competent summaries, ordinary marketing copy, predictable service journalism, and endless variations of information that already exists.

Human excellence can survive. Human originality can survive. Strong reporting, distinctive voices, earned authority, and meaningful relationships with readers can survive.

But they will not survive merely because we label them human.

They will survive because they are better.

I do not believe AI is the end of writing or publishing. I believe it is the end of several comfortable assumptions about what made our work valuable.

That may be frightening, but it can also be clarifying.

Our advantage is not that we can assemble sentences. Machines can now do that.

Our advantage is that we have lived, learned, failed, remembered, questioned, and cared. Our responsibility is to bring those qualities to our work, use these new tools honestly, protect the rights of creators, and remain accountable to our readers.

This is my manifesto. You are free to disagree with it. In fact, I expect some of you will.

But the industry cannot afford to ignore what is happening.

The press is already running. We can still decide what belongs on it.