Table of Contents
Journalism has a dirty little secret. Artificial intelligence is already helping write, edit, research and produce the news Americans read every day. Reporters are using it. Editors are experimenting with it. Major news organizations are quietly incorporating it into their daily operations. And, increasingly, journalists are beginning to acknowledge what much of the profession seemed reluctant to admit even a year or two ago.
AI isn't coming to journalism. It's already here. The question is whether the profession has the courage to acknowledge that reality, establish meaningful standards and stop treating an extraordinary technological advancement as something respectable journalists must pretend doesn't exist.
After nearly 50 years in this business, much of it covering sports and community news in northeastern Minnesota, I've witnessed plenty of technological revolutions. We moved from typewriters to computers, from darkrooms to digital photography, from printing presses to websites capable of publishing breaking news in seconds. Every transformation produced anxiety, resistance and predictions that journalism as we knew it was finished.
Yet nothing quite compares with artificial intelligence, largely because previous advancements changed how journalists delivered their work. AI has the potential to change how that work is produced.
The numbers are remarkable.
Muck Rack's 2026 State of Journalism report found that 82% of surveyed journalists had used at least one AI tool during the previous 12 months, up from 77% in 2025. ChatGPT was used by 47%, while substantial numbers reported using AI transcription tools, Grammarly and other applications.
A separate Reuters Institute for the Study of Journalism report, published in November 2025, found that 56% of 1,004 British journalists surveyed were using AI professionally at least once a week. Another 27% used it less frequently. Only 16% reported never using it for journalistic tasks.
These are not insignificant numbers buried in an obscure technology publication. They describe a profession undergoing a fundamental transformation, whether its practitioners are comfortable discussing it publicly or not.
And the technology is moving well beyond automated transcription and spellchecking.
The Reuters Institute found that 22% of surveyed journalists used AI for story research at least monthly, 16% for brainstorming, 16% for generating portions of articles, such as headlines, and 10% for generating first drafts.
That final number deserves attention. We are no longer talking exclusively about machines checking punctuation or transcribing interviews. Some working journalists are already asking AI to help construct the stories that eventually appear beneath human bylines.
To be clear, those figures do not mean AI is independently reporting the news or that most journalists have abandoned traditional writing. They do establish that the boundaries between human writing and machine assistance are changing rapidly.
Even the venerable Associated Press has recognized the transformation.
In July, AP updated its newsroom standards to permit specific uses of artificial intelligence, including preliminary research, document summaries, transcription, translation, headline suggestions and copy editing. The organization emphasized that verification, editorial judgment and accountability remain the responsibility of its journalists, with disclosure standards when AI plays a material role in published content.
Think about that for a moment. One of the world's most respected journalism organizations, whose standards have helped define professional newswriting for generations, now formally recognizes AI as a legitimate reporting and editing assistant under controlled circumstances.
That's not an endorsement of machine-produced journalism. It's an acknowledgment that the technology can serve legitimate editorial purposes without surrendering the principles that make journalism credible.
And that distinction is everything.
I've never believed the quality of journalism should be measured by how painfully or inefficiently an article was produced. The reader doesn't care whether a reporter wrote a story on a battered typewriter, a laptop or a smartphone. The reader wants accuracy, clarity, fairness, perspective and information worth knowing.
If an experienced journalist can use AI to digest a 200-page government report, organize complicated financial information, identify unanswered questions or improve a poorly structured article, why should that assistance be considered professionally unacceptable?
I would argue that refusing to consider such tools merely to preserve an outdated image of journalistic purity makes little sense.
But there is another side to this argument, and it deserves even greater scrutiny.
Artificial intelligence can produce convincing sentences without knowing whether those sentences are true. It can invent sources, manufacture quotations, misunderstand statistics and present fundamentally inaccurate conclusions with astonishing confidence. An inexperienced or careless journalist could easily turn those weaknesses into a published disaster.
And an unscrupulous publisher could use AI to flood a community with inexpensive articles that look like journalism but contain little original reporting, local knowledge or independent verification.
That possibility concerns me far more than an experienced reporter using AI to improve a legitimate piece of journalism.
Consider local news, where the financial condition of many traditional news organizations has already weakened the profession. Newsrooms have lost reporters, experienced editors and institutional memory. Important public meetings receive less coverage. Government decisions can escape meaningful scrutiny simply because nobody has the time or resources to investigate them.
AI can help a small newsroom operate more efficiently. It cannot responsibly serve as an excuse to eliminate the reporters whose work makes that newsroom relevant.
A computer program cannot replace a journalist's firsthand understanding of Duluth politics, the history of a contentious school board decision or the relationships developed over decades covering a community. It cannot independently establish what happened behind closed doors, cultivate the trust of reluctant sources or accept professional responsibility when its published information proves wrong.
It can assemble words about a Minnesota Duluth hockey game. It cannot substitute for the reporter who watched the game, questioned the coach, understood a player's development and recognized the significance of a performance that might otherwise go unnoticed.
Those are not minor distinctions. They are the foundation of credible reporting.
I also believe journalists must stop pretending that using AI and practicing legitimate journalism are automatically incompatible.
There is nothing inherently unethical about using technology to organize notes, review grammar, compare documents or strengthen a narrative. There is something profoundly unethical about publishing invented quotations, presenting unverified claims as facts, concealing substantial machine-generated work or allowing a computer to masquerade as an independent reporter.
The standards should be unmistakable: Every fact must be verified. Every quotation must be authentic. Confidential sources and sensitive unpublished information must be protected. Original reporting must remain original reporting. And readers deserve disclosure when AI substantially contributes to the content they are being asked to trust.
The responsibility belongs to the journalist whose name appears above the article and the editor or publisher responsible for putting it before the public.
There is another uncomfortable reality. Some of the loudest opposition to AI has become tangled with an entirely legitimate fear that publishers will use technology to reduce payroll expenses. Journalists have good reason to worry about their livelihoods. News organizations have already demonstrated a willingness to sacrifice reporting capacity in pursuit of financial survival.
The Reuters Institute survey found that 62% of British journalists considered AI a major threat to journalism. Their concerns should not be dismissed as technological ignorance or professional stubbornness.
But we should distinguish between the technology itself and the decisions made by the people who control it.
A publisher who replaces experienced reporters with inexpensive machine-generated content is making a business decision. A journalist who uses AI to improve research efficiency while maintaining rigorous reporting standards is making a professional decision. Those choices are not equivalent.
Indeed, thoughtful AI assistance could become particularly valuable to independent journalists and small publishers who lack the research departments, copy desks and production staffs available to larger organizations.
For a one-person news operation, the technology offers something genuinely significant: additional editorial capacity without sacrificing independence. Used carefully, it can make ambitious, detailed journalism more practical. Used recklessly, it can destroy the credibility upon which an independent publication depends.
This brings us to the question many journalists still seem reluctant to confront. Should we admit using AI? Absolutely.
Not every routine spelling correction or automated transcription requires a warning label. But when AI materially influences the preparation or composition of a published article, readers have a legitimate interest in understanding its role.
Disclosure should be straightforward, proportional and free of technological theatrics. Readers shouldn't have to guess whether they're receiving original reporting, machine-assisted writing or fully automated content.
There should be no professional shame in acknowledging responsible use of an editorial assistant, whether that assistance comes from another human being or a sophisticated software application.
What deserves criticism is deception, not efficiency.
After five decades in journalism, I've become convinced that the profession's long-term survival will depend less on defending traditional methods than on defending traditional standards.
Accuracy. Independence. Fairness. Accountability. Skepticism. The obligation to seek facts rather than manufacture them. The willingness to challenge powerful institutions and explain complicated issues without distorting the truth.
Those principles were essential when reporters carried notebooks and worked against afternoon printing deadlines. They remain essential in an era when artificial intelligence can generate a thousand polished words in seconds.
The technology is advancing faster than the profession's willingness to discuss it openly. That gap creates confusion, suspicion and unnecessary distrust at precisely the moment journalism can least afford another credibility problem.
We should close it.
I have no interest in romanticizing the old newsroom simply because I spent most of my professional life working in one. Nor am I interested in surrendering the craft I've spent nearly half a century learning to a machine that cannot be held accountable for what it publishes.
There is a sensible, defensible position between those extremes.
Embrace the assistance. Verify the information. Protect the reporting. Respect the reader. Be transparent about how the work was produced.vAnd never confuse a well-written article with a well-reported one.
Artificial intelligence may become the most powerful editorial assistant journalism has ever known. It may also become one of the most dangerous instruments ever placed in the hands of careless publishers.
The difference will be determined by the people who use it.
The future of journalism should not be a choice between human reporters and artificial intelligence. It should be about experienced journalists using every legitimate tool available to produce stronger, more accurate and more valuable work.
The byline still matters. The truth still matters. And the journalist behind both must always matter most.
