Investigative journalism is hard, expensive, time-consuming, and often dangerous. I’ve been sued, jailed in Cuba for spying, and even kidnapped over my 40-year career.

But AI is turning the tables and putting new powers in my hands. I’d go so far as to say it’s democratising investigative journalism by giving those powers to anyone.

It’s a big claim, so let me break it down in a literal field report and reveal how I’m doing the accountability reporting of an entire newsroom with a laptop and a $30-a-month subscription. AI has been crucial in how I was able to build a 19,000-strong newsletter audience in 99 countries, which led to me being invited to address the UN and advise multiple governments… in less than three years.

Let’s begin by busting a myth. This isn’t getting the AI to write 1,700 versions of the same article and blasting it across the socials. Nope, this is the opposite. Using the AI to do the grunt work, freeing me up to do the rest.

Used the right way, AI shifts the asymmetry in publishers’ favour. For decades, Big Tech has sent armies of lawyers, comms teams, and lobbyists to control the narrative. Journalists have been left to fight with notebooks and deadline pressure. The information gap was a moat, and tech knew it.

But now, it’s draining. A reporter with the right prompts can now process documents at a speed that used to require billion-dollar resources. Journalists can now do the digging they were trained for, and use the tech to turn it into hard-hitting reporting.

Now a single journalist can hold Google or Meta, Iran or Russia, history or political doublespeak to account—and still have time for lunch.

Finding the needle

I have sources like journalists always have, but many of mine are no longer human. I have alerts set up in search, and notifications on court papers and SEC filings. My tipoffs come in as a steady stream 24 hours a day. Many are nothing, but then one is a trigger. It happened the other week.

The judge in charge of breaking Google’s search monopoly ruled that a technical committee must be established to do the job. Everyone wants to know who they are.

An alert popped up on my phone from an obscure automated court reporting AI that the three had been named. The link gave me the court papers. I was off.

I dropped the committee members a line on LinkedIn, used AI to research their careers and found years-old articles that hinted at their personalities. Within an hour, I profiled them and broke the story. Then I sat down to write it. Boom. Job done.

An antitrust lawyer whose name had been linked with the role rang me from New York saying he’d scoured the court papers and couldn’t find the names. They were buried, I told him. Deep. Even with all his resources he couldn’t find it with his team. Now we’re connected too.

Financial forensics

Apples and pears. That’s journalists and accountants. Journos do words and geeks do maths. Only I love both, because you need to follow the money to find the truth.

Every quarter, the tech firms report their earnings to the US Securities and Exchange Commission. They are interminable—dull and data-laden, but full of gold. Best of all, the tech titans who love nothing more than privacy have to put these numbers out in the open to satisfy their obligations to shareholders.

It’s a goddam turkey shoot. I have uploaded years of financial filings, shareholder updates, Wall Street earnings calls into my own small language model. I’ve trained it on my historic reporting, so it knows what I am looking for. For example, is the 90.5%* of advertising Google sends to its own search and YouTube rising or falling? Spoiler: It’s always rising.

That’s why the open web is in danger of collapsing. It’s why publishers have no money to fund newsrooms. It’s also why experienced single operators like me can strike.

Apple’s numbers told me it’s reliant on China amid a tariff war. Meta’s told me 97.6% of its income relies on ads. Snap showed it relies on selling ads to the youngest teens.

These data points that justify headlines are often buried in footnotes, YOY comparisons that used to take weeks and geeks to reveal—but not now.

The fact-checking machine

AI’s a brilliant sub-editor after you’ve taught it your style.

My SLM—I call it RoboRicky—then reads the draft and alerts me if I’ve forgotten a relevant fact in a previous article. It even suggests charts and images. My newsletter contains more than three million words of investigative journalism now, but there’s more than 10 million words of source material in RoboRicky.

It checks every word against the source material to confirm it’s accurate and flags anything that it thinks is wrong.

I’ve also had fun using Google’s Gemini to punch holes in its CEO’s fibs, and had OpenAI run an analysis on whether its deals to buy content were fair. (They weren’t.)

RoboRicky + my brain + my instincts + a superhumanly unwise amount of coffee now power my one-man newsroom. No team. No budget. Me, a laptop, and my killer AI pal.

*Footnote: RoboRicky corrected an error. I’d said Google sent 89% of its ads to search and YouTube. The actual number from the Q4 SEC filing last month (and not reported but calculable by doing the complex maths) was even more, 90.5%.

The post AI is making the one-man newsroom a reality appeared first on The Media Copilot.

For local news organizations stretched across dozens or hundreds of communities, the old model of listening to a single police scanner in a single newsroom no longer scales. Readers still expect their local outlet to be first on breaking stories—crime, fires, severe weather—but reporters and editors cannot monitor every frequency, Facebook group, and traffic camera on their own.

Dataminr, a real-time breaking news detection platform, was built to fill that gap. By aggregating information from police scanners, traffic cameras, social media posts, government advisories, corporate disclosures and other public sources, its AI flags early signs of news events and delivers them as geographically filtered alerts.

1. Turning a firehose of public data into usable alerts

Dataminr’s distinguishing feature is breadth. The company says its systems perform trillions of computations daily across billions of data points from more than one million public sources. Those inputs range from emergency radio traffic and public sensor data to posts on mainstream and alternative social media platforms.

For editors, that volume is useless without filtration. Dataminr’s value lies in narrowing the firehose to a manageable stream tailored to a newsroom’s geography and interests. Users define coverage areas by city, county, or region and set topic parameters for crime, safety, weather, infrastructure and other beats.

Once configured, the platform delivers alerts tagged by severity—Flash for major national stories, Urgent for regional breaking news, and standard alerts for lower-priority items. Editors like Patch’s national breaking news editor, Anna Schier, rely primarily on Urgent alerts as a balance between comprehensiveness and noise.

2. Early warning in unfamiliar markets

For reporters working deeply in a single town, a text from a trusted source at city hall or a tip from a community Facebook group may still be the most valuable signal. But for regional or national desks responsible for many communities, those relationships are harder to maintain.

Dataminr is designed for that second scenario. The platform is most effective when covering unfamiliar territory—places where a newsroom has an audience but no permanent presence. It can surface reports of heavy police presence, highway closures, severe storms or industrial fires in areas that would otherwise be invisible until much later.

In practice, that head start often amounts to minutes rather than hours. But in breaking news, minutes matter. The platform’s own materials note that alerts may arrive five minutes to several hours before a story would surface through more traditional means such as social browsing or official press releases.

3. Flexible integrations that match newsroom habits

Dataminr’s alerts can be delivered through multiple channels: email, a web dashboard, Slack or Microsoft Teams, and mobile push notifications. Each method supports a different part of the workflow.

• • Email provides a searchable archive of past alerts, useful for following up on tips that didn’t initially appear significant.

• • The web dashboard offers real-time monitoring, with maps and filters suited to editors on active shifts.

• • Team messaging tools distribute alerts to groups in seconds, reducing the lag between detection and action.

• • Mobile notifications extend coverage beyond office hours, particularly for weekend and overnight shifts.

The platform’s implementation guide emphasizes that its effectiveness depends less on technology than on process: assigning clear responsibility for monitoring, defining escalation paths, and aligning alert settings with actual coverage capacity.

4. Support for distributed staffing models

Patch.com’s use of Dataminr illustrates one of the platform’s core strengths: enabling central editors to support local reporters across a wide footprint. With one reporter often covering an entire community, regional editors and breaking news leads need tools to watch for major developments when local staff are away or occupied.

Dataminr’s geographic filters let those editors monitor multiple markets simultaneously. When an Urgent alert appears from a town without an on-duty reporter, they can decide whether to publish a brief, hold for confirmation, or assign the story to a nearby editor.

Over time, this capability helps maintain a consistent standard of responsiveness across a network, even as staffing levels and experience vary from market to market.

5. A complement to, not replacement for, newsroom sourcing

Dataminr does not replace the work of cultivating local sources. Its own case study materials emphasize that the platform works best “in tandem” with relationships built by on-the-ground reporters.

Editors interviewed about the tool stress that they treat Dataminr alerts as starting points. Official sources, such as law enforcement or government accounts, may justify quick, clearly attributed briefs. Alerts that originate from unverified social posts or vague scanner traffic require additional verification before publication.

The company’s Multi-Modal Fusion AI is designed to cross-verify events across multiple data types, on the assumption that genuine incidents leave multiple signals. But the system cannot eliminate the need for human judgment about what constitutes a story and when information is reliable enough to share.

Who should consider Dataminr

Based on the available documentation, Dataminr fits best for:

• • Newsrooms covering many markets, particularly at regional or national scale

• • Operations with dedicated staff who can monitor real-time alerts during key hours

• • Organizations that compete on being first to report breaking events

• • Teams willing to invest in verification workflows and staff training

The platform is less well suited to single-community outlets with strong local sourcing, very small newsrooms (fewer than five staff), or organizations that primarily need social media trend analysis rather than breaking news detection.

Newsrooms interested in Dataminr can request demos and pricing by contacting [email protected]. Initial setup typically takes one to two hours, with full customization and training over one to two weeks.

Frequently Asked Questions

The post Why multi-market newsrooms choose Dataminr for breaking news detection appeared first on The Media Copilot.

Anna Schier’s first job in media involved one piece of equipment and a lot of patience. Sitting in a small Chicago newsroom, she listened to a police scanner with “one ear on the scanner at all times,” waiting for the crackle that signaled news.

Today, as a national breaking news editor at Patch.com, Schier still listens for emergencies. But instead of a single radio tuned to one city, she monitors Dataminr, an AI-powered platform that scans thousands of public sources—from police and fire channels to social media posts and traffic cameras—to flag early signs of news events in communities she has never set foot in.

For a network of hyperlocal sites that covers more than 1,900 communities with lean staffs, that shift has changed how Patch finds and prioritizes breaking stories. It has also raised new questions about verification, information overload, and what it means to cover a place you know mostly through data.

A thinly stretched local network

Patch Media runs local news sites in markets ranging from Los Angeles to Wheatland, Wyoming, population just under 4,000. The stories swing from nationally relevant—major crimes, severe weather—to the deeply local: stolen e-bikes, school board fights, town parades.

Most of that reporting still happens the old-fashioned way. Editors and reporters attend municipal meetings, file from courthouses, and rely on tips from residents and community Facebook groups. In many small markets, a single reporter covers everything.

Schier and other breaking news editors sit above that structure, providing backup when big stories break or when a local editor is on vacation. “If we have a story that’s breaking in an unstaffed area that might still be of interest to readers or someone’s on vacation or we just need a set of extra hands, that’s kind of where I come in,” she says.

The model creates a central problem: how to know, quickly, when something newsworthy is happening in a town where the company has no one on the ground.

Replacing one scanner with thousands of feeds

Dataminr is designed for that gap. The platform ingests information from police scanners, traffic cameras, government advisories, corporate disclosures, blogs, alternative social media services, and public sensors such as power outage and flight data. Its AI runs trillions of computations on billions of inputs each day, looking for patterns that suggest something is happening.

For newsrooms, those patterns arrive as alerts filtered by geography and severity. Patch configures its Dataminr feed to focus on the kinds of stories its readers expect: crime, fire, weather, and significant community events.

Alerts fall into three broad categories:

• Flash: major national stories, such as presidential announcements or large-scale disasters
• Urgent: regional breaking news including serious crimes, accidents, and severe weather
• Alert: lower-priority items

Most Patch editors rely on Urgent alerts, which capture local breaking news without overwhelming staff. The system pipes notifications into email inboxes for searchability and into Dataminr’s web dashboard for real-time monitoring during shifts.

Instead of a single voice crackling over a scanner, Schier now sees a stream of structured alerts, each with enough context—location, source type, initial details—to decide whether to assign a story, flag it for a local editor, or watch for further confirmation.

Speed as a survival strategy

Patch’s audience expects immediacy. “When a story breaks, get one sentence up… and then build it out from there,” Schier says of the company’s approach to high-urgency stories. “The most important thing is to let people know as quickly as possible.”

Without Dataminr, that speed is hard to sustain across 1,900 markets. Local editors know their beats, but they can’t monitor every possible signal alone. For breaking news editors covering unfamiliar territories, the challenge is heavier: there are no entrenched source networks to lean on.

Dataminr doesn’t eliminate that problem, but it narrows the gap. The platform can surface an alert about heavy police presence, a highway closure, or an emerging wildfire in a matter of minutes. In some cases, Dataminr surfaces information five minutes to several hours before it would otherwise land on an editor’s radar via social media or a local tip.

For a newsroom competing with television, radio, and social feeds in each market, that head start is often the difference between leading coverage and playing catch-up.

verification-remains-a-human-job”>Verification remains a human job

The volume and variety of Dataminr’s inputs create their own risks. The platform flags everything from official police bulletins to unverified social media posts, and it does not—and cannot—guarantee that every alert is accurate.

“Dataminr’s job is to raise alarm bells and let me decide what to do with them,” Schier says. “So I don’t necessarily expect that it’s going to be right and I don’t ever trust that it’s right. I always look at the source of where it’s coming from first.”

Patch treats Dataminr alerts as starting points, not publishable facts. Official sources, such as law enforcement accounts, may justify a short initial story with clear attribution, followed by updates. Alerts based on more ambiguous signals—multiple eyewitness posts, scanner chatter without confirmation—require phone calls, cross-checking, and often patience.

Dataminr’s own Multi-Modal Fusion AI is designed to reduce false alarms by looking for corroboration across data types. “A real breaking news event is likely to have corroboration across multiple data sources,” says Mike D’Orio, the company’s chief product officer. Even so, the burden of judgment rests with editors.

Managing overload and tailoring the feed

One of the most consistent challenges noted in Dataminr’s own documentation is volume. Even with filtering, the platform can surface more potential stories than a small team can handle.

Patch’s editors respond by tightening geographic and topical filters and relying heavily on mapping features to see where clusters of alerts are emerging. “Use the filters, use the mapping feature,” Schier advises. “These kinds of tools work the best when you personalize them to meet your needs and to align with your goals.”

Dataminr’s implementation guidance echoes that approach: start with restrictive settings and expand gradually. Newsrooms are encouraged to:

• Define primary coverage areas by county or city clusters
• Designate secondary markets for occasional coverage
• Assign dedicated staff to monitor alerts during peak hours
• Create email rules to sort alerts by priority and desk

Without that discipline, Dataminr can feel less like a scanner and more like a firehose.

Where Dataminr fits—and where it doesn’t

Dataminr is not a universal solution. Its strengths lie in:

• Multi-market coverage: Newsrooms that cover wide regions or national beats, especially in unfamiliar communities
• Early warning: Detecting events before they surface through traditional channels
• Backup capacity: Enabling central editors to support local staff during absences or major events

The platform is less critical when a single reporter has deep, longstanding ties in a small community. In those cases, a text from a parent at a school board meeting or a call from a trusted source may still beat any algorithm.

Subscription costs are also a factor. Dataminr offers custom pricing based on organization size and includes unlimited newsroom licenses, but its own documentation acknowledges that newsrooms with fewer than five staff members may struggle to justify the expense.

Alternatives fill other parts of the monitoring stack. NewsWhip focuses on social media trending rather than raw breaking alerts. Rolli emphasizes tracking disinformation and connecting journalists with vetted experts. Traditional scanner apps provide direct access to emergency communications but require constant attention and only cover a limited geography.

A new kind of listening

For Schier, the move from a single police scanner to Dataminr has not changed the core of the job so much as it has extended its reach.

“Nothing is going to replace the work that a local reporter has done to be informed about a community, to build relationships,” she says. “But Dataminr can be used in tandem with that to get you the story a little bit faster.”

In a fragmented local news landscape, where a small staff may be responsible for dozens of towns, that “little bit faster” can be the difference between being the place readers go first—or not at all.

Newsrooms interested in Dataminr can contact the company’s news team at [email protected]. Implementation typically takes one to two hours for initial setup and one to two weeks for full customization and training.

The post Inside Patch’s AI-era listening post: how Dataminr rewired its breaking news workflow appeared first on The Media Copilot.

For editors responsible for covering dozens of communities at once, the appeal of Dataminr is obvious. The platform claims to process vast amounts of public information—from police scanners and traffic cameras to social media posts and power outage sensors—and turn them into early alerts about fires, crashes, protests and other potential stories.

But entrusting a breaking news workflow to an algorithm raises practical and ethical questions. How reliable are the alerts? What kinds of data is the system ingesting? And what responsibilities do newsrooms retain when they rely on a third party to tell them where to look?

Available case studies and implementation guidance offer a partial picture.

Risks identified in Dataminr’s use for newsrooms

Dataminr works by aggregating and analyzing public information, not by providing official confirmation. That distinction matters. The platform flags what it believes may be newsworthy based on patterns across sources, including social media posts that could be incomplete, inaccurate or intentionally misleading.

Editors interviewed about the tool stress that they do not treat alerts as facts. “Dataminr’s job is to raise alarm bells and let me decide what to do with them,” says Patch.com‘s national breaking news editor Anna Schier. “So I don’t necessarily expect that it’s going to be right and I don’t ever trust that it’s right. I always look at the source of where it’s coming from first.”

Relying on Dataminr without robust verification workflows could lead to premature publication of unverified claims—particularly under the pressure to be first on breaking events. Newsrooms using the platform must guard against that temptation.

Another risk is information overload. Even with geographic and topical filters, Dataminr can produce more alerts than small teams can handle. Without clear triage protocols, staff may miss important signals amid lower-priority noise.

Finally, because Dataminr monitors public social media and other open sources, its output may reflect the biases and blind spots of those platforms. Events in communities with less online activity may be underrepresented, while incidents that generate viral posts may be overemphasized.

Controls and practices that mitigate those risks

Dataminr’s documentation and spokespersons describe several technical approaches intended to improve reliability. The company’s Multi-Modal Fusion AI cross-references signals across data types, on the theory that genuine breaking events will generate multiple independent traces—a scanner transmission, social posts, perhaps sensor data—while false alarms may not.

In practice, the most effective safeguards appear to be editorial rather than algorithmic. Newsrooms are advised to:

• Treat alerts as tips rather than publishable information
• Differentiate by source type, publishing faster when alerts come from official accounts and more cautiously when they originate from social chatter
• Build verification checklists for different alert categories, including calls to local officials, cross-checks against other monitoring tools, and on-the-ground confirmation when possible
• Define responsibility for monitoring and response on each shift, so alerts don’t fall into a gap between desks

Dataminr itself does not store journalists’ private source information or reporting, according to available materials. It surfaces activity already visible in public information streams.

Security and privacy considerations

The Dataminr newsroom documentation reviewed focuses more on workflow and use cases than on technical security architecture. Specific details about data storage, encryption, access controls and retention policies are not provided in the source materials.

Given the nature of the platform—continuous monitoring of public information and location-based alerting—newsrooms should:

• Consult their legal teams about how Dataminr collects and processes social media content and other public data
• Clarify whether any newsroom-specific information (such as user configurations or alert histories) is stored and how it is protected
• Ensure that no internal, non-public data is inadvertently fed into the system

Because Dataminr works with public sources, the primary privacy questions revolve around platform design and vendor practices rather than the newsroom’s own audience data. Even so, organizations that have adopted strong privacy positions may wish to understand how Dataminr’s business model and partnerships intersect with their own commitments.

A tool, not a gatekeeper

For all its automation, Dataminr does not absolve newsrooms of responsibility. Its strongest use cases—early warning in unfamiliar markets, backup coverage when local staff are offline—are also the ones where verification is hardest and mistakes can carry the greatest consequences.

Editors who have integrated the platform into their work emphasize that it is most effective when tightly configured and paired with human judgment. “Nothing is going to replace the work that a local reporter has done to be informed about a community, to build relationships,” Schier says. “But Dataminr can be used in tandem with that to get you the story a little bit faster.”

News organizations considering Dataminr should approach it as a powerful but fallible signal generator. The platform can widen a newsroom’s field of vision and buy precious minutes in fast-moving situations. It cannot decide what is newsworthy, what is true, or what is safe to publish.

Those decisions remain, appropriately, in human hands.

Dataminr’s news team can be reached at [email protected] for organizations seeking detailed security and privacy documentation beyond what is available in public case studies.

Frequently Asked Questions

The post Can you trust Dataminr with your breaking news workflow? appeared first on The Media Copilot.

Patch.com’s readers expect their local site to be first on big stories, whether it’s a highway closure, a neighborhood fire or a fast-moving storm. But with one reporter often covering an entire town, and editors responsible for clusters of markets, the company needed a way to see beyond a single police scanner or a handful of Facebook groups.

Dataminr, a real-time breaking news detection platform, has become one of Patch’s central tools for doing that work at scale. By scanning thousands of public sources and flagging potential news events, the system gives editors minutes—or sometimes hours—of advance warning they would otherwise struggle to get.

The gist

Dataminr acts as a digital scanner for Patch’s distributed newsroom.

• AI-driven alerts flag breaking events across more than 1,900 communities
• Geographic and severity filters keep the volume manageable for small teams
• Editors treat alerts as tips, not facts, and verify before publishing

How they use it

Patch’s editors have built Dataminr into their daily and overnight routines.

• Geographic filters: Coverage areas are defined by city clusters and regions, ensuring alerts match Patch’s footprint.
• Alert tiers: Editors rely primarily on Urgent-level alerts for serious local events such as crimes, accidents and severe weather.
• Multi-channel delivery: Alerts arrive via email for searchability and through a web dashboard for real-time monitoring during shifts.
• Backup coverage: Breaking news editors use alerts to step into unstaffed markets, posting initial briefs and then coordinating with local reporters.

Key numbers

Dataminr and Patch do not publish specific performance metrics, but the platform’s documentation notes:

• Coverage scale: Patch operates in more than 1,900 U.S. communities
• Setup time: Initial configuration typically takes 1–2 hours; full customization and training 1–2 weeks
• Detection advantage: Alerts can surface events 5 minutes to several hours before traditional discovery methods

What to watch for

Patch’s experience with Dataminr underscores the need for guardrails.

• Verification burden: Alerts based on social media or scanner traffic require confirmation before publication.
• Information overload: Even with filtering, smaller teams can be overwhelmed without clear protocols.
• Fit by newsroom size: Dataminr’s own guidance suggests that very small outlets may struggle to justify the subscription cost.
• Workflow dependence: The platform delivers value only when someone is actively monitoring and empowered to act.

For Patch, Dataminr has not replaced reporters’ local relationships. It has, however, given editors a broader view of where trouble is starting—and a better chance of staying ahead of it.

Newsrooms can contact [email protected] for demos and tailored pricing.

The post How Patch uses Dataminr to keep its breaking news edge appeared first on The Media Copilot.