Editorial Policy on the Use of Artificial Intelligence

The Southern Maryland Chronicle uses artificial intelligence as a research and production aid in select stages of our editorial process. AI does not report, decide, or publish on its own. Every article we publish is written, reviewed, and approved by a human editor before it reaches our readers. This page explains, in plain terms, where and how AI is involved in our work.

Our Standard: Human-Led, AI-Assisted

Artificial intelligence at the Chronicle is a tool, not a byline. It does not replace reporting, editorial judgment, or accountability. A member of our editorial staff reviews, fact-checks, and approves every piece of content — text, headline, graphic, and social post — before it is published. If AI touches a story at any stage, a human has the final say on what runs.

Our Proprietary 4-Agent Workflow

Newsroom research and drafting support is organized through a proprietary four-agent workflow. Each agent handles a distinct function, and each agent’s output is treated as raw material for a human editor — not a finished product. A reporter or editor reviews, verifies, and assembles the final article using both their own reporting and the agents’ output combined.

• Agent 1 — Research & Fact-Checking: gathers background information, source material, and context on a topic, and checks factual claims against available records.
• Agent 2 — Angles, Risk, Structure & Style: reviews story angles, flags legal or editorial risk, checks word count, structure, grammar, and adherence to AP Style.
• Agent 3 — SEO & Social Media: prepares search and social optimization elements for the article (see below for specifics).
• Agent 4 — Quality & Prompt Compliance: reviews the work of the other agents against the original assignment to confirm nothing was missed or left incomplete.

The human writer or editor then assembles the published article, drawing on their own original reporting alongside what the AI agents surfaced. No article is published directly from an AI agent’s output.

AI-Assisted Graphics

AI tools may be used to help produce graphics for our stories. All AI-generated graphics follow a strict internal style prompt designed to produce clean, professional visuals consistent with our brand — not the oversaturated, uncanny look often associated with unedited “AI slop.” Every graphic is reviewed by a human before use, and photojournalism, and images purporting to depict real events, are never AI-generated.

AI in SEO & Social Media

AI plays its largest supporting role in helping our content reach readers, not in shaping what the story says. Here is how it’s used:

Search Engine Optimization (SEO)

• Keyword and search-intent research to identify what readers are actively searching for
• Suggested headline and subheadline variations for clarity and searchability
• Draft meta titles and meta descriptions for review and editing
• Internal linking suggestions to related Chronicle coverage
• Readability and structure checks (subheads, paragraph length, scannability)
• Structured data / schema suggestions to help search engines correctly categorize articles
• Identifying trending or under-covered local topics worth assigning

Social Media

• Draft social post copy tailored to each platform’s format and tone
• Suggested hashtags and post timing
• Repurposing an article into a social thread, carousel, or short-form summary
• Draft alt text and image captions for accessibility
• Headline/copy variant testing to see what resonates with readers

All SEO and social output is reviewed and edited by our staff before publishing or posting. AI suggests; our editors decide.

Our newsletter is produced with the assistance of Claude AI, under the same human-review standard applied to every other AI-assisted product.

What We Don’t Do

The Chronicle never uses AI to artificially amplify its own work — we do not use bots, fake engagement, automated boosting, or any other artificial means to inflate our reach on social media or in search results. Our reach is earned through real reporting and real readers.

We do not pay for, authorize, or condone the misuse of AI with our copyrighted work. This includes unauthorized scraping, reproduction, or training of AI models on Chronicle content without our consent.

A Note on AI-Detection Tools

A number of programs now claim to be able to tell whether a piece of text was written by a human or by AI. Readers, platforms, and aggregators increasingly rely on these tools, so we want to be direct about their limits: independent testing has repeatedly found these detectors to be highly unreliable, with significant false-positive rates — flagging ordinary human writing, including clean, well-edited, or simply structured prose, as “AI-generated” when it is not.

Because the Chronicle is transparent about using AI as a research and production aid (as described above), a published article may, at times, be flagged by one of these tools. That flag is not evidence of anything. It does not mean a story was written by AI instead of a human, and it does not override the standard described on this page: every article we publish is written, reviewed, and approved by a human editor. We’d encourage readers and other outlets not to treat the output of these detection programs as fact.

The Tools We Use

Our workflow draws on multiple commercial AI models, matched to the task at hand: Claude AI, Google Gemini, OpenAI’s ChatGPT, and xAI’s Grok. These tools are independent, general-purpose AI services — they are not integrated into, or connected with, the Chronicle’s internal systems, subscriber data, or publishing platform. No AI vendor has access to our internal editorial systems.

Our Commitment

We believe readers deserve to know how their news is made. As our use of AI evolves, we will keep this page updated to reflect our current practices. If you have questions about this policy or believe an error in our coverage may be related to AI-assisted research, please contact our editorial team.

Southern Maryland Chronicle — Editorial Policy August 2026