Building an AI-Driven Local Newspaper: Lessons from Tucson
AI-generated, human-reviewed.
How Nicholas de Leon Is Using AI to Build a Local Newsroom for Less Than $40 a Month
On this episode of Intelligent Machines, journalist Nicholas de Leon revealed how he’s been able to single-handedly create, automate, and publish a comprehensive local news service using affordable AI tools and just a few hundred lines of code. His innovative approach to the Tucson Daily Brief not only automates the coverage of local government and public events, but also demonstrates a new model for sustaining essential community journalism in the AI era.
How AI Is Powering Daily Local News Coverage
Nicholas de Leon’s Tucson Daily Brief is a bespoke, local news automation project serving the Tucson, Arizona area. By combining a range of AI models, transcription engines, and custom scripts, he delivers daily reports on city council meetings, public safety updates, new business openings, and more—most of which would otherwise go unreported as legacy local newsrooms continue to shrink.
According to de Leon on this week's episode, the site pulls public data and aggregates stories from dozens of local sources. What’s unique is that it goes beyond simple curation: using large language models (LLMs) and real-time automated transcription, it generates original news stories from raw public meeting feeds and official documents, often beating traditional outlets to publication.
The Tech Stack: Affordable Tools & Automation
The backbone of Tucson Daily Brief is a collection of affordable, consumer-accessible AI models. De Leon starts with aggregation—using AI to synthesize key updates from local TV, radio, and newspaper feeds. But the real shift comes from his "Live AI Reporter" system, which:
- Transcribes city government meetings automatically in real time using Deepgram and FFmpeg.
- Feeds those transcripts to an LLM (Anthropic’s Sonnet model) to generate coherent, journalistic write-ups highlighting the most impactful news from hours-long public meetings.
- Adds a brief human editorial review to correct name pronunciations and fact-check for accuracy before publishing.
Beyond written news, de Leon clones his voice with ElevenLabs to create automated daily podcasts, updates the site’s newsletter, and even generates local crossword puzzles tied to current Tucson headlines—all using “glue code” and cloud-based APIs.
Impact: Local News, Hyper-Efficient and Accessible
As highlighted on Intelligent Machines, this approach means one reporter can cover four municipalities’ worth of local business every day, before breakfast. De Leon’s AI-driven process surfaces liquor license filings, business permits, and public safety alerts that might otherwise be neglected.
His goal isn’t world domination—it's to have "a seat at the table" and keep the local public informed in a news desert. Notably, his system frequently produces the first (and sometimes only) reports on city events, offering timely access to information that matters most for Tucson residents.
Expanding Beyond News: Sports Simulations and Job Tools
De Leon’s use of AI goes further. He’s developed automated sports simulations (Deep Dugout and MatchDayInference), experimenting with AI-driven baseball and World Cup reports. He’s also launched MyNextAssignment, a job-matching tool for journalists and media professionals that scrapes hundreds of listings, rewrites resumes, and drafts cover letters with minimal human input.
What You Need to Know
Nicholas de Leon’s Tucson Daily Brief covers local news entirely with affordable AI and minimal manual work.
Automates city council meeting write-ups, leveraging AI-powered transcription and summarization.
Publishes daily podcasts using a cloned voice, produces local crosswords, and manages social media—all AI-assisted.
Operates for roughly $40/month in cloud AI costs, showing that robust local journalism can be both efficient and cost-effective.
Human review is critical for accuracy and cultural context, especially with names and sensitive topics.
Approach could provide a model for other small communities or single-person newsrooms coping with the decline in legacy media.
The Bottom Line
The Tucson Daily Brief demonstrates that a single motivated journalist, using commodity AI models and smart automation, can revive and even expand local news coverage at a fraction of historical costs. As Nicholas de Leon explained on Intelligent Machines, this method is not only scalable but also immediately relevant for communities facing a vacuum in trustworthy civic news. The hybrid model—AI for speed and scale, human for judgment and final edit—points toward a sustainable future for local news creation in the digital age.
Subscribe to keep up with the latest technology and AI journalism trends:
https://twit.tv/shows/intelligent-machines/episodes/883
