by Beryl Lipton And Cooper Quintin

This article was first published by the Electronic Frontier Foundation (EFF) on September 10, 2021. It is reprinted here with permission.

Prison telecommunication companies have historically been in a voice-based business with one way to profit: charging exorbitant rates for phone calls.

The business model has paid off for both the companies and their partners, the jails and prisons, but in recent years, at the behest of prison reform advocates and the families of people who are incarcerated, the Federal Communication Commission (FCC) has set a cap on per-minute charges on interstate phone calls.

In the meantime, prison communication companies—Global Tel*Link, Securus, and others—have been considering fresh features and potential revenue streams, building on all those voices that go through their lines every day. Prisoners’ calls are already being monitored and recorded (both legally and not so legally). What if something could be done with all that data?

In the name of security and fraud prevention, prison communication companies have imagined a few ways to store, analyze, and use the trove of voices, as articulated in patents on voice printing, matching, and recognition filed in recent years. Importantly, patents often precede the actual development or deployment of a technology. Though applications may demonstrate an interest in advancing a particular technology, these intentions don’t always progress beyond the proposal, and many inventions that are described in patent applications don’t wind up being built. What we can glean from a patent application is that the company is thinking about the technology and that it might be coming down the pipeline.

Securus and GTL have both previously made clear that there’s a future worth fighting for in inmate phone surveillance. As early as 2010, GTL noted in an application for a call management and monitoring system that it would be “highly beneficial for communication systems in penal institutions to incorporate biometrics as an additional security device.” GTL later claimed Securus violated this patent and the two spent multiple years suing each other before reaching a confidential settlement in 2016.

Since then, both have continued to file patents for voice-based security features, particularly those based on cross-referencing databases of voice samples with other information.

One GTL patent suggests that voice matching could be to verify and authorize callers, facilitating the automated disconnection of any caller deemed “unauthorized” or with too few funds to afford the call.

Another GTL patent describes how multi-modal audio mining could monitor conversations and map networks of phone call participants as part of an automated monitoring system built on speech recognition and analysis. The mining would include “topic detection,” keyword searches, and pattern analysis across multiple conversations from different facilities. According to the patent, being able to compare conversations as they happen to words and phrases that may not be obvious to a monitoring guard could improve security in a facility.

“[T]he topic detection process matches phrases contained in an external topic database, such as a slang database. One advantage of using a slang database is that it provides a repository of slang utterances that may be made during a telephone conversation including, but not limited to, prison slang, gang slang, radical religious slang, terrorism related slang, slang for weapons, violence, and threats of violence, and drug related slang….For instance, for the slang term “blow” a major topic might be drugs and a minor topic might be drug dealing. By detecting topics discussed during a monitored telephone conversation by analyzing at least the transcript of the monitored telephone conversation, a user, such as a correctional officer, may query the multimedia data warehouse for telephone conversations involving discussions of, for example, drags in general and drug dealing in particular without having to know the exact phrase or phrases used during the conversation.

Voice printing and mining systems are already in place throughout the country, including in Arkansas, Texas, and Florida.

The companies’ statements and contracts with facilities show that at least some of their ideas were ultimately implemented and that the reach of their voice printing technology is capable of expanding facilities’ collection of biometric data beyond prisoners and into the greater carceral community too.

GTL, for example, has marketed a collection of features designed to record and store information from all parties: “identify the location of a mobile device that has accepted a call” from a facility, “detect and identify words or a sequence of words spoken by the inmate and called party,” and measure emotion throughout the call, and others.

GTL describes its Data IQ as a “suite of biometric applications” that can “isolate all inmate “touch points” with the outside—telephone calls, video visitation sessions, in-person visitation sessions, commissary deposits, financial transactions, and more.” Its Voice IQ, GTL claims, can “record calls in distinct channels so that both the inmate and called party sides of the conversation are available for review together or separately.”

Florida is among the states that has access to GTL’s branded Call IQ system, according to a contract signed between GTL and the Florida Department of Corrections in 2020.

If the claims are true, the technology could identify and profile anyone who has a voice that crosses into a prison, including all the parents, children, lovers, and friends of incarcerated people.

Securus has also considered the possibilities of voice databases. In a patent published in January 2021, a process is described for collecting audio samples of individuals’ voices both at the moment of intake and while inmates are communicating with people on the outside.

Facilities are known to acquire voice samples by threatening a loss of privileges should an inmate refuse to bow to the surveillance state. “Here’s another part of myself that I had to give away again in this prison system,” one inmate recalled in a 2019 article by The Intercept after he was told that failing to help train the system to recognize his voice would result in a loss of his ability to use the phone.

Another Securus patent for “Multi-party conversation analyzer and logger” highlights the use of voice printing in order to facilitate “the investigation of networks of criminals, by gathering associations between phone numbers, the names of persons reached at those phone numbers, and voice print data.” The company claims that the system could be used to identify ex-inmates calling possibly accomplices in prison, a sort of association mapping that could be used on anyone making calls to the prison for any reason.

In a statement sent to EFF after publication of this series, Aventiv, the parent company of Securus, confirmed that it would be reviewing its open patents for alignment with its new organizational goals, writing:

“We at Aventiv are committed to protecting the civil liberties of all those who use our products…. Our organization is focused on better serving justice-involved people by making our products more accessible and affordable, investing in free educational and reentry programming, and taking more opportunities—just like this one—to listen to consumers.”

We at EFF will be following up with Aventiv to learn more about the results of that review. Notwithstanding Aventiv’s statement, Securus’s marketing materials still pitch its voice recognition surveillance tool to prison officials. Sold as part of the company’s Investigator Pro service, Securus claims that “You’ve Never Seen Voice Biometrics Like This.”

Other companies have also been considering correctional call monitoring. Confinement Telephony Technology was recently awarded a patent for monitoring inmate calls, and CenturyLink filed a patent for a self-learning corrections call monitoring system.

Regardless of whether the voice printing technology described in the patents above is implemented or which company develops it, major privacy and data security questions need addressing:

• What happens to people’s biometric data once facilities collect it?

• Where is the data stored?

• Will it be shared with anyone else?

• Who can access the data?

Given the invasiveness of any biometric collection system and the many unanswered data privacy issues they create, it’s crucial that the public push jails and prisons to limit adoption of these technologies. 

Subscribe today

Already a subscriber? Login