This is the age of the citizen journalist where people on the street armed with a smartphone and a Twitter account are often the first on the scene of an accident, fire or other emergency. These folks know about an incident that requires emergency services well before first responders.

Dataminr introduced a new product on stage at TechCrunch Disrupt today in New York City that searches the Twitter firehose for emergency situations throughout the city, and channels news alerts to first responders.

The system is designed to filter out fake announcements, jokesters, pranksters and other non events, so first responders aren’t wasting their time chasing rumors of non-events.

Company CEO, Ted Bailey was on stage today along with Ben Krakauer, Assistant Commissioner for Strategy and Program Development at NYC Emergency Management, and Masha Gindler from the Mayor’s Office on Digital Strategy discussing how they could make use of tech like this to improve emergency response.

“Dataminr has built the first technology that can identify breaking information in social media. Many other companies detect trends or analytics. We don’t. We detect that first nugget of information when it first gets published. Ultimately when we [issue an] alert, it is a headline, an alarm bell to first responders to look here and investigate via other sources. A social media signal is never going to be definitive. It is just the most performative early trip wire for breaking information…” Bailey explained.

The idea is that Twitter sends many signals about events happening throughout the city and Dataminr can provide the city’s emergency services with additional layers of information about the nature of the situation. As Krakauer explained this is not really about replacing 911 so much as augmenting it.

“We want you to call 911. The answer is not to send a tweet, but anecdotally during the largest emergencies — crane collapses, building explosions, large fires — at the same time we have information coming in from 911 calls, we’re also seeing information come in from the social media stream,” Krakauer explained.

While the project began in part as a public service, it has grown into a $3 million contract over three years (although Krakauer pointed out the city might not spend anywhere close to that). Bailey said it also offers his company a way to expand the news alert platform to other first responder organizations across the country (and presumably the world).

As for the organized spread of false information that we’ve seen over the last few years with false reports of a chemical spill in Louisiana, an Ebola outbreak and police shootings in Atlanta that never actually happened, Bailey asserts that when there is a false report, it soon becomes apparent enough and his company’s algorithms are tuned to pick up when a report is real and when it’s not.

“Social media far more powerful verifier than spreader of false information” Bailey told TechCrunch. That’s because the signals that come in during an actual emergency including photos, provide a clear indication when something is happening (or if it’s fake).

Dataminr got in some trouble last year when it reportedly began pitching a more robust product for monitoring by law enforcement personnel. There was a huge social media backlash and Twitter eventually limited law enforcement access. To further muddy the waters in that instance, Twitter is an investor in Dataminr, but Bailey said today that the company was never providing surveillance information for law enforcement as had been asserted.

“We still provide a news alert product for law enforcement for first response, what never happened was a surveillance product to any law enforcement. It never happened,” Bailey said.

Dataminr was founded in 2009 and has raised over $183 million. The platform also offers news alerts for news organizations, financial services and private corporations.

Note: This article originally stated in error that “Dataminr firehose access had been pulled for law enforcement purposes.” It has been updated.