The landscape of journalism is undergoing a profound transformation as independent reporters, navigating the complexities of a post-newsroom environment, increasingly turn to advanced AI tools like Anthropic’s Claude to streamline their workflows and amplify their output. This evolution is not about replacing human journalists but rather about augmenting their capabilities, allowing them to focus on core journalistic strengths like source cultivation and original reporting. Technology reporter Alex Heath, who transitioned to an independent Substack newsletter last year, exemplifies this trend. Heath now utilizes Claude, an AI-powered conversational agent, to generate initial drafts of his articles, dramatically reducing the time spent on the often-arduous "zero-to-one" phase of writing.
Heath’s innovative approach involves integrating Claude Cowork, a sophisticated AI tool, directly into his journalistic process. This integration extends to his digital ecosystem, connecting with his Gmail, Google Calendar, Granola AI transcription service, and Notion notes. A critical component of this setup is a meticulously crafted custom "skill" for Claude. This skill comprises a detailed set of instructions, including what Heath refers to as the "10 commandments" of writing in his distinctive style. This bespoke training data includes his previous articles, specific guidance on newsletter structure, and detailed notes on his unique voice and writing cadence.
By feeding Claude this comprehensive profile, Heath effectively trains the AI to mimic his journalistic persona and stylistic nuances. The AI agent then automates the initial drafting, a process that previously consumed significant cognitive energy. Following Claude’s first draft, Heath engages in an iterative refinement process, collaborating with the AI for up to 30 minutes, suggesting revisions and steering the narrative. While Heath still contributes significantly to the final product, this AI-assisted workflow has resulted in a substantial time saving, reportedly reducing his writing time by 30 to 40 percent weekly. "I’ve always hated the zero-to-one process of writing a story," Heath remarked. "Now, it’s actually kind of fun. Going out on my own, I realized I need AI to help with the volume."
The Rise of AI in Independent Journalism
Heath’s adoption of AI is emblematic of a broader movement within the tech reporting community and beyond. For independent journalists who have opted out of traditional newsroom structures, the loss of built-in resources such as editors and fact-checkers can be a significant hurdle. Instead of simply using general-purpose AI tools like ChatGPT for basic content generation, these independent reporters are actively working to recreate these essential support functions using AI. This strategic implementation raises important questions about the evolving role and perceived value of human journalists in an increasingly automated media landscape.
The Human Element in an AI-Driven Era
The pervasive integration of AI into writing, editing, and even fact-checking processes prompts a critical examination of what unique contributions human journalists bring to the table. A recent study by Google DeepMind researchers shed light on this dynamic, suggesting that a passive or "lazy" approach to AI utilization can lead to more homogenous, less creative, and less distinctive writing. The study indicated a tendency for AI-generated content, when not carefully guided, to adopt a more neutral stance, potentially diminishing the unique voice and opinion that readers often seek. To effectively leverage AI, journalists must possess a deep understanding of the fundamental reasons why audiences engage with their work in the first place. This necessitates a clear articulation of the value proposition beyond mere information delivery.
For many, like Alex Heath, the core value lies not solely in the prose but in the journalistic enterprise itself: the ability to secure exclusive information and provide timely insights. Claude, in his case, acts as a force multiplier, enabling him to dedicate more time to cultivating sources and delivering crucial information to his subscribers. This symbiotic relationship with AI echoes a historical precedent in journalism: the rewrite desk. In bygone eras, reporters in the field would transmit their findings via phone to writers in the newsroom, who would then craft these raw details into polished articles for publication. In essence, Claude has become Heath’s modern-day rewrite desk, facilitating a division of labor that allows him to excel at the reporting aspect of his profession. "I feel like I’m cheating in a way that feels amazing," Heath admitted. "I never did this because I liked being a writer. I like reporting, learning new things, having an edge, and telling people things that will make them feel smart six months from now."
AI as an Editorial Partner: Jasmine Sun’s Approach
Jasmine Sun, a former product manager at Substack and now an independent newsletter author focusing on AI and Silicon Valley culture, offers another compelling perspective on AI’s role in journalism. While Sun’s own research, highlighted in a recent Atlantic article, suggests that AI models can struggle with creative writing due to limitations in their post-training capabilities, she has found significant promise in Claude as an editorial assistant.
Sun’s strategy involves feeding Claude past articles and notes on her writing style, but with a specific directive: the AI’s sole purpose is to enhance and refine her existing voice and taste, explicitly avoiding sycophancy. She has instructed Claude that it "should never write a sentence for her. Your goal is to elicit out of Jasmine by providing feedback." This nuanced approach emphasizes collaboration rather than delegation.
Sun shared a segment of her instructions to Claude, underscoring this distinction: "You are not a co-writer. You cannot perceive—you don’t have experiences, sources, scenes, or emotions to draw from. Your role is to help Jasmine write like the best version of herself—not just who she is on the page now, but who she’s trying to become as a writer. That means understanding both her current voice and her aspirations, including the writers and qualities she’s reaching toward."
When asked if she ever felt tempted to let Claude do the heavy lifting, Sun responded, "I think [Claude] forces me to work harder than I would otherwise. With a human editor, they’re calling you on your bullshit. They’re not letting you get away with lazy reporting or floppy prose."
The use of AI in editorial roles has not been without its critics. Sun recounted receiving criticism from individuals who felt that AI could not replicate the transformative power and challenging dialogue provided by a human editor. However, Sun argues that for many Substack creators, the cost of a human editor is prohibitive. By integrating Claude and instructing it to be a rigorous critic, she has, in her view, made her editorial process more robust. "To me, it is kind of like using Grammarly," she explained. "You have a tool that says, ‘this sentence is bad,’ and then I have to go fix it. [Claude is] more willing to tell me this entire section is bad and you should cut it. It’s like a higher level of thinking and abstraction than Grammarly can do."
Redefining Value in the Age of AI: Casey Newton’s Perspective
Casey Newton, the author of the influential newsletter Platformer, believes that the advent of sophisticated AI necessitates a re-evaluation of a publication’s core value proposition. "I think there’s an interesting distinction here," Newton stated. "If the value is in the information, not the writing, then I think people will care less that AI did most of the writing. If the value is in voice and opinion and argument and analysis, it seems like it’s cheap to use AI to do the whole thing."
Newton, who has recently focused on news analysis, acknowledges that the increasing capabilities of AI are prompting him to shift his editorial strategy. "I actually need to shift the balance," he shared. "I need to do less news analysis and more original reporting."
While Newton is not currently employing AI for writing Platformer, he has been inspired by Sun’s AI editor concept and has begun experimenting with creating a Claude agent trained on his own articles to replicate a similar editorial feedback loop. "I’ve just been really struck by it," Newton remarked. "At its best, its feedback is about as good as the feedback I’ve gotten from human editors."
Navigating Sensitive Reporting and the Craft of Writing: Taylor Lorenz
Taylor Lorenz, author of the User Mag Substack, utilizes AI for various aspects of her media business, employing Gemini for SEO-optimized YouTube descriptions and Claude for data analysis. However, Lorenz maintains a firm boundary regarding the use of AI in her writing and editing processes. She expresses a lack of trust in current AI systems when dealing with sensitive reporting materials and finds that AI has not yet proven its utility for the nuanced craft of writing and editing.
"I am a journalist because I like to help people understand the world and bring light to different issues," Lorenz stated. "I don’t want the AI to do that." Her commitment to the human element in journalism underscores a belief that certain core journalistic functions, particularly those involving deep investigative work and empathetic storytelling, remain intrinsically human endeavors.
AI as a Book Production Accelerator: Kevin Roose’s Experimentation
Kevin Roose, a technology columnist for The New York Times, is leveraging AI tools to expedite the production of his forthcoming book on the race to build artificial intelligence. He estimates that AI has accelerated the process by two to three years.
Most recently, Roose has assembled a team of Claude agents to assist in editing his book, headed by a "Master Editor" agent. This AI ensemble includes sub-agents responsible for fact-checking, ensuring stylistic consistency with Roose’s voice, and providing constructive feedback. It is important to note that Roose continues to collaborate with human editors on the project.
Despite the advancements, Roose has not relinquished the primary writing responsibilities to AI. Similar to Sun and Newton, he believes his own writing abilities currently surpass those of AI models. "I think the models tend to be fairly generic and depersonalized, but also, I like doing this," he commented.
Roose identifies as an AI optimist, anticipating that AI models will eventually surpass human capabilities in most domains. However, he emphasizes the enduring advantage of being human, at least for the present. "I am not under some romantic illusion that I possess a special, irreplaceable perspective. But what I am is a person, and I think that for now people, at least some people, like hearing from people," he concluded. This sentiment highlights a crucial aspect of contemporary journalism: the human connection and the inherent value audiences place on narratives and insights delivered by fellow humans.