After bulletins are issued, the software can use artificial intelligence to analyze multiple reports, which are often publicly available, across cities, regions, or even multiple states.
The AI looks for patterns, such as the use of a similar vehicle or a breaking and entering technique, then alerts users if it links crimes from multiple reports (each department decides whether and how far afield its bulletins will be included in the AI search). The AI also tries to display relevant bulletins to each user, such as alerting a robbery detective investigating one break-in of similar prior incidents.
“If you’re inside of a big [law enforcement] agency like a Boston or a Seattle, the left hand often doesn’t know what the right hand is doing,” said Matt White, cofounder and chief executive, in an interview at the firm’s offices near Teele Square in Somerville. “We’ve taken what is a kind of an ad hoc process around the country and turned it into something that is a searchable, usable database.”
In one case, White said the company’s software helped law enforcement agencies in two different states connect the same domestic terrorist group to incidents in which weather radars were destroyed. (Multitude Insights declined to disclose details of the incidents.)
In another case, the system connected a string of thefts by a credit card scammer in California, thanks to a baseball cap the crook was wearing, White said.
In a demonstration at Multitude’s headquarters, White and chief technology officer Frank Conroy showed how a law enforcement officer could rapidly create a bulletin by selecting categories such as the type of crime from drop-down lists, fill in more details, and upload photos or videos. At the bottom of the report, other officers can leave comments or tips.
Police departments using Multitude’s software contacted by the Globe either declined to comment or did not respond.
With cofounder Akihiko Izu, a Japanese lawyer who White met at MIT’s Sloan School, Multitude’s team has raised more than $5 million from investors including New York-based venture capital firm Commonweal Ventures.
At a time when police use of AI has come under scrutiny, the startup’s software does not include the capability to recognize faces in crime photos or videos included in the digital bulletins.
AI facial recognition has been found to be biased in some studies and is prohibited for use by police in some areas. “It’s really hard to deploy nationally anything having to do with facial recognition, so until the policy is a little bit more clear, that’s been a business decision to stay out of that,” White said.
The current mess of paper and emailed bulletins often leads to information overload and ignored reports, according to Dean Esserman, former chief of police in New Haven. And when a suspect is caught for one crime, officers rarely have time to research whether they might have been involved in other cases.
“It would be great if you had the time and no pressure to keep plodding along, trying to link other cases, other hints, other pieces of evidence or leads,” said Esserman, who now advises Multitude. “This software can help that happen.”
Before attending MIT as part of a joint program with Harvard’s Kennedy school, White spent more than eight years in the Navy, crammed into the back of aircraft flying intelligence-gathering missions. He came away impressed with the armed forces’ technology for analyzing intel. Then on a ride-along with Boston police during grad school, he sensed an opportunity to modernize the antiquated bulletin system.
“I spent a lot of time trying to find the signal in the noise,” White said. “What’s actually important is helping an officer get to the next important thing, not just the next thing.”
Aaron Pressman can be reached at aaron.pressman@globe.com. Follow him @ampressman.