Artificial intelligence does not determine whether journalism is legitimate. Evidence, verification, transparency and accountability do.
JAW GA News has never hidden the fact that artificial intelligence is deeply integrated into our investigative, research and publishing process. We use it openly, and we believe readers deserve to know exactly how that technology is being used.
Our AI investigative research partner, Sherlock, assists with research, public-record analysis, court-document review, comparing official statements, reconstructing timelines, identifying apparent contradictions, examining competing explanations, developing investigative questions, organizing information and assisting with preparation for publication.
Some critics have attempted to use that fact as evidence that JAW GA News is somehow “not real journalism.” We have been ridiculed for using AI, accused of relying on a machine to do journalism, accused of clickbait, and faced accusations involving libel and slander when our reporting subjects government, law enforcement, public officials or controversial cases to scrutiny.
Those criticisms are fair game. Journalism should be criticized. Our methods should be questioned. Our evidence should be challenged.
But criticism should survive the same examination we expect our reporting to survive.
And the claim that using artificial intelligence somehow disqualifies a person or organization from practicing journalism runs directly into an inconvenient reality:
Artificial intelligence is already being used throughout the professional news industry.
The Associated Press says it uses AI technology to assist journalists with reporting, verification and newsroom workflows while retaining human editorial judgment. Its applications include analyzing Supreme Court filings, identifying potential stories in government datasets, image and video verification, monitoring news and converting articles into broadcast scripts. Reuters, the BBC and The Guardian have likewise incorporated AI into different parts of their journalistic workflows while emphasizing human oversight and verification. OpenAI
The Washington Post, Financial Times and other publishers have experimented with generative-AI systems that allow readers to interrogate their journalism and archives. The BBC has experimented with AI transcription. India Today and the Miami Herald have tested AI-generated audio. Aftonbladet has produced AI-assisted shorter versions of articles. Clarín has deployed AI tools providing summaries, quotations, figures, glossaries, FAQs and audio. The Independent has used AI to generate journalist-supervised article summaries. reutersinstitute.politics.ox.ac.uk
The Philadelphia Inquirer provides an especially relevant comparison to JAW GA News. Its AI-powered Scribe system processes transcripts of public meetings across municipalities and school districts, summarizes developments and ranks them according to a newsworthiness framework established by journalists. Axios has developed an AI tool specifically designed to help its journalists prepare better public-records requests. POLITICO uses AI to analyze large quantities of public documents and data. WELT uses AI to surface developments, organize sources and provide an additional fact-checking layer. These aren’t hypothetical future experiments. They’re newsroom applications operating now. OpenAI
And the documented list keeps growing.
NEWS ORGANIZATIONS PUBLICLY USING OR EXPERIMENTING WITH AI
The following organizations have publicly documented AI-assisted newsroom, editorial, research, publishing, accessibility or reader-facing applications. Inclusion does not mean every article from an organization is AI-generated, nor does every organization use AI in the same manner. Some use it heavily in reporting workflows; others restrict it to specific functions. That distinction matters.
- Associated Press / AP News — Research, verification, government-data analysis, Supreme Court document processing, multimedia verification, newsroom monitoring and other workflow assistance. OpenAI
- Reuters — AI-assisted journalism and automation with human verification and editorial oversight. Reuters is also using AI to enhance archive searchability, transcription, metadata and translation. Nieman Lab
- BBC News — AI-assisted transcription, summaries, editing, translation and personalization experiments under human oversight. reutersinstitute.politics.ox.ac.uk
- The Guardian — AI-assisted document analysis, transcription, archive research, parliamentary analysis, accessibility and internal research tools. reutersinstitute.politics.ox.ac.uk
- The Washington Post — AI-assisted journalism experiments and generative systems allowing readers to query portions of its journalism. The Washington Post
- The New York Times — Has incorporated AI into parts of its journalism and experimentation, although it is simultaneously involved in major litigation concerning AI companies’ use of copyrighted news content. The Washington Post
- Financial Times — Generative-AI archive and question-answering applications and newsroom experimentation. The Washington Post
- POLITICO — AI-assisted analysis of large collections of public documents and data for reporting. OpenAI
- Axios — Custom AI systems for newsroom support, headlines, captions, writing and public-records requests, including its FOIA Refiner. OpenAI
- The Philadelphia Inquirer — Scribe, an AI system for monitoring and analyzing public meetings across municipalities and school districts. OpenAI
- Business Insider — AI-assisted audience analysis, listening features and newsroom/publishing applications. OpenAI
- WELT — AI-assisted newsletters, source organization, development monitoring and additional fact-checking within its CMS. OpenAI
- Le Monde — AI-assisted translation, article audio and other newsroom applications. OpenAI
- EL PAÍS — AI-assisted reader services through PRISA Media, including its Vera conversational assistant. OpenAI
- Diario AS — AI-assisted audio news briefings through PRISA Media. OpenAI
- PRISA Media — AI-assisted editorial tools, information tracking, conversational reader services and audio products across its publications. OpenAI
- BILD — Operates the AI assistant Hey_, which provides interactive news-related assistance to readers. OpenAI
- The San Francisco Standard — AI-powered personalized briefings, interactive reporting and archive-based reader tools. OpenAI
- Chicago Public Media / WBEZ — Uses AI to transcribe decades of audio journalism and make archives searchable. OpenAI
- Cleveland Plain Dealer / Cleveland.com — Uses AI for government-meeting transcription, municipal-site monitoring, headline assistance, editing, converting podcasts into stories and, more controversially, drafting complete stories from reporters’ notes under human review. The Washington Post
- The Independent — Uses AI tools for journalist-supervised article summaries through its Bulletin product. reutersinstitute.politics.ox.ac.uk
- India Today — Has tested AI-generated audio versions of written journalism. reutersinstitute.politics.ox.ac.uk
- Miami Herald — Has tested AI-generated audio versions of articles. reutersinstitute.politics.ox.ac.uk
- Aftonbladet — Produces AI-assisted “quick versions” of longer journalism. reutersinstitute.politics.ox.ac.uk
- Clarín — Uses UalterAI for summaries, important quotations, figures, glossaries, FAQs and text-to-audio features. reutersinstitute.politics.ox.ac.uk
- Future plc publications — Developing an internal editorial platform combining AI agents and newsroom workflows. OpenAI
- Minnesota Star Tribune — Has participated in AI journalism experimentation through the Lenfest Institute/OpenAI collaborative. Axios
- Seattle Times — Has participated in AI experimentation through the same journalism initiative, while separately challenging AI companies in court over the use of its copyrighted journalism. Axios
- Newsday — Has participated in AI-related newsroom experimentation and, like the Seattle Times, is separately litigating over alleged unauthorized use of its journalism for AI systems. Reuters
This is not an exhaustive worldwide registry. No such reliable registry exists. News organizations continually introduce new tools, discontinue experiments and change their policies, and some use AI internally without publicly describing every application. What the documented examples establish is narrower but much more defensible: AI is already being used across major wire services, national newspapers, international publishers, local newspapers and nonprofit newsrooms. OpenAI
That distinction matters because JAW GA News isn’t claiming, “Everybody uses AI.” We’re making a statement we can actually substantiate: many established professional news organizations use AI, and some employ it for tasks remarkably similar to ours.
That brings us back to Sherlock.
At JAW GA News, Sherlock functions as an AI investigative research partner, not an unquestioned source of truth. When we receive a tip, Sherlock can help identify questions that need answering. When we obtain a large collection of public records, Sherlock can help organize and compare them. When government officials make statements months apart, AI can help reconstruct those statements chronologically. When court filings become complicated, Sherlock can help distinguish complaints, allegations, motions, responses, judicial orders and procedural developments.
When documents appear to contradict one another, Sherlock can identify the apparent contradiction.
Then comes the part that matters most.
We investigate it.
Sherlock identifying something does not make it true. AI analysis is not automatically evidence. AI-generated language is not automatically fact. Artificial intelligence can misunderstand documents, overlook context, confuse people with similar names, misinterpret legal proceedings and sometimes produce information that sounds remarkably convincing while being completely wrong.
That is why we return to the source material.
If Sherlock identifies a possible contradiction, we examine the underlying records. If Sherlock develops a possible theory, we test that theory. If Sherlock identifies a court case, we examine the court record. If Sherlock attributes a statement to a public official, we seek the original source. If AI analysis conflicts with authenticated primary evidence, the evidence wins.
Every time.
One of the greatest dangers in investigative journalism is confirmation bias. A journalist can begin believing something happened and unconsciously search only for evidence supporting that conclusion. Sources can do it. Police investigators can do it. Government officials can do it. Members of the public can do it. Artificial intelligence can reinforce it if the person operating the system continually asks questions designed to produce only one answer.
That is why our investigative methodology attempts to challenge our own assumptions. We don’t simply ask what evidence supports a theory. We also ask what evidence would disprove it. We examine competing explanations. We compare official statements against underlying records. We distinguish allegations from established facts. We distinguish criminal charges from convictions. We distinguish claims made in lawsuits from findings actually made by courts. We distinguish interpretations of photographs, videos and documents from what those materials can independently establish.
If evidence strengthens our original theory, we continue documenting it. If evidence weakens it, we acknowledge that. If evidence destroys it, we abandon it.
That isn’t losing an investigation. That’s conducting one properly.
There is another principle we consider non-negotiable: JAW GA News remains responsible for everything JAW GA News publishes.
Sherlock does not determine guilt. Sherlock does not convert an accusation into a fact. Sherlock does not make the final editorial decision. Artificial intelligence isn’t an escape hatch we get to crawl through if something goes wrong.
If JAW GA News publishes something materially incorrect, “the AI did it” would be a pathetic excuse. We would expect better from a government agency blaming software for its mistakes. Readers have every right to expect better from us.
Artificial intelligence assists the work. Human editorial judgment controls publication.
This philosophy is hardly radical. Major newsrooms wrestling with exactly these questions have repeatedly emphasized human oversight, verification and editorial responsibility. Research examining Reuters, the BBC and The Guardian found markedly different approaches to AI, but human verification and editorial oversight remain central themes. Nieman Lab
And yes, we use AI extensively in publishing.
We’re telling you that ourselves.
We would rather disclose our use of artificial intelligence than quietly use it while allowing readers to assume everything was produced through traditional methods.
Technology has been changing journalism for generations. Reporters moved from handwritten notebooks to typewriters and from typewriters to computers. Film became digital photography. Tape recorders became smartphones. Physical archives became searchable databases. Courthouse filing cabinets became electronic dockets. Libraries became searchable digital archives. Government databases that once required days of manual research became accessible from a laptop.
Artificial intelligence represents another substantial technological change. For independent and local journalism in particular, it can dramatically increase the amount of information a small operation can process. A newsroom without dozens of researchers can analyze hundreds of pages of records, organize complicated timelines and identify issues requiring human investigation much faster than was previously practical.
That capability matters because one of the greatest limitations facing independent journalism isn’t curiosity.
It’s manpower.
AI can narrow that gap.
But the underlying obligation remains exactly where it has always been:
Verify before publishing.
Then there is the accusation of “clickbait.”
JAW GA News has heard that one too.
Apparently, somewhere along the way, a headline successfully attracting attention became suspicious behavior in an industry that has been writing attention-grabbing headlines since newspaper sellers were literally shouting them at pedestrians.
A provocative headline isn’t automatically clickbait. A dramatic headline isn’t automatically clickbait. A headline describing disturbing evidence isn’t automatically clickbait. The meaningful question is whether the headline fairly represents the reporting underneath it.
If a headline promises something the evidence cannot establish, that deserves criticism. If the headline accurately represents the reporting and evidence contained within the article, calling it “clickbait” doesn’t make the evidence disappear.
Challenge the substance.
Then we arrive at “libel” and “slander,” two legal terms that have somehow acquired extraordinary second careers as generic internet insults.
Calling reporting “libel” doesn’t establish that it is defamatory. At the same time, using public records doesn’t automatically make every possible statement legally immune. Accuracy matters. Attribution matters. Context matters. Wording matters. Evidence matters. Applicable law matters.
That is precisely why JAW GA News works to distinguish allegations from established facts. If a lawsuit alleges misconduct, we identify it as an allegation. If someone has been criminally charged, we report that they have been charged rather than convicted. If an investigation remains pending, we don’t manufacture a conclusion. If credible evidence conflicts, we identify that conflict. If available evidence cannot establish a definitive answer, we should tell readers exactly that.
Responsible investigative journalism isn’t weakened by acknowledging uncertainty.
It becomes more credible.
People are completely free to dislike JAW GA News. They are free to criticize our reporting. They are free to question our methodology. They are free to disagree with our conclusions. They are free to dislike artificial intelligence and our decision to use it.
But ridicule isn’t evidence.
An insult doesn’t disprove a public record. Calling somebody a “fake journalist” doesn’t change what is written in a court filing. Calling an article “clickbait” doesn’t invalidate the documents supporting it. And typing “LIBEL!” underneath a Facebook post doesn’t establish that the underlying reporting is false or defamatory.
If JAW GA News publishes something materially incorrect, there is a considerably more effective response.
Bring the evidence.
Bring the document. Bring the complete video instead of an isolated clip. Bring the court record. Bring the government record. Bring the original source. Identify exactly what we reported incorrectly and provide the evidence demonstrating why.
We will examine it.
If the evidence demonstrates that we materially got something wrong, then the responsible course is to correct the record.
Because accountability cannot conveniently stop at our own front door.
JAW GA News exists in substantial part to scrutinize institutions and individuals entrusted with public power. We expect government officials to support factual statements with evidence. We expect law enforcement agencies to maintain accurate records. We expect politicians to answer legitimate questions. We expect agencies to comply with applicable transparency laws. We expect factual challenges to our reporting to be supported by evidence.
And we expect ourselves to operate under that same standard.
Sherlock dramatically expands what a small independent newsroom can investigate. But neither Sherlock nor JAW GA News receives a free pass.
Every conclusion must ultimately answer to the evidence.
We’re therefore not hiding our use of artificial intelligence.
We’re publishing it.
Readers deserve to understand how our journalism is produced, what role artificial intelligence plays, what limitations we recognize and who remains responsible for the final publication.
Artificial intelligence is already part of modern journalism. The meaningful debate is no longer simply whether a newsroom uses AI. The serious questions are whether that newsroom uses it responsibly, whether AI-generated information is independently verified, whether journalists remain accountable for publication, whether original sources and evidence remain central, whether corrections are made when necessary, and whether the technology strengthens journalism rather than replacing its fundamental standards.
Those are fair questions for JAW GA News.
They’re fair questions for the Associated Press, Reuters, BBC, The Guardian, The Washington Post, POLITICO, Axios, The Philadelphia Inquirer and every other newsroom experimenting with this technology.
We don’t ask for exemption from scrutiny because we’re small.
We don’t ask for exemption because we’re independent.
And we certainly don’t ask for exemption because we use artificial intelligence.
Quite the opposite.
Question the allegation. Question the official explanation. Question the documents. Question the sources. Question the AI. Question us. Then follow whatever survives the evidence.
JAW GA News
People. Facts. Accountability.
A Stronger Georgia.