A ShotSpotter review center seen in a screen capture from a company video.

Somerville is being called on cancel its contract for ShotSpotter gunfire detection by the ACLU Massachusetts in a press release Thursday.

It’s “surveillance technology that markets itself as a solution to gun violence but instead is unreliable, ineffective and poses serious threats to basic civil rights,” said the letter to the City Council signed by three members of the local ACLU’s Technology for Liberty Program.

Contributing to the press release, city councilor at large Willie Burnley Jr. said that “based on the testimony of experts and statements the council has received from residents, I have deep concerns about the reliability of ShotSpotter and its potential to create unsafe situations in our community. Rather than taking a reactionary approach to public safety, I’d rather our city use the federal funds we receive to proactively invest in programs that allow us to intervene in crises before they escalate to the point of violence.”

While ShotSpotter, a product by the company SoundThinking, advertises a 97 percent accuracy rate and a 0.5 percent false positive rate, the ACLU alleges that this is misleading. ShotSpotter operates based on an algorithm that gets data from several microphones (the recommendation is 15 to 20 per square mile), but human analysts are also available to “reclassify” erroneous reports.

In a document obtained by The Associated Press, it is reported that as many as 10 percent of ShotSpotter alerts are reclassified by analysts. The sounds that the system has erroneously identified as gunshots have included cars backfiring, fireworks, dumpsters, church bells and construction.

Following the release of leaked ShotSpotter location data in February, there have been increased calls to abandon the technology. At a City Council meeting in Cambridge, which also uses ShotSpotter, representatives from the ACLU and Digital Fourth raised concerns about the Cambridge Police Department’s use of the technology. Digital Fourth, a local civil liberties group with an interest in privacy and surveillance, alleged that ShotSpotter data released by police has been simply reproduced since 2021, pointing out a persistent typo as evidence. The ACLU also presented concerns about the effectiveness of the technology and stated its intent to participate in a Public Safety Committee hearing April 2 to comment further on the issue.

ShotSpotter is in use in several Massachusetts cities, including Boston, Worcester and Springfield. Fall River was the fifth Massachusetts city to adopt the technology, but abandoned it in 2017 after reporting a 41 percent error rate and several other activations resulting in no evidence or arrests. ShotSpotter also missed or failed to report several gunfire incidents in neighborhoods where sensors were installed.

In Cambridge, neither of the two confirmed shots-fired incidents between Jan. 1 and Feb. 26 were reported by ShotSpotter, according to a February police report called BridgeStat.

There were six confirmed shootings or shots-fired incidents last year, excluding the Jan. 4 shooting of Arif Sayed Faisal by Cambridge police, four of which were reported by ShotSpotter. Gunfire that killed 27-year-old Danasia Greene in November was one of the two incidents ShotSpotter did not report, despite witnesses reporting hearing six to eight gunshots and a sensor allegedly being placed in the area.