For the last three years, while studying at the University of Washington, Yael Eiger was also teaching in local prisons. She headed to a penitentiary in Monroe weekly, as well as one in Gig Harbor, working with organizations such as the Prison Math Project and the Freedom Education Project of Puget Sound.
Doing this work, Eiger, a doctoral student in the Paul G. Allen School of Computer Science & Engineering, noticed a disconnect. The technology incarcerated people had to use to communicate with anyone outside of prison was “really inefficient, unusable, frustrating and expensive,” Eiger said. For instance, each message cost money to send, but it didn’t necessarily reach the recipient. Why was this technology so much worse than the messages she sent for free? And how did its failings affect the people using it?
In three recent papers, Eiger and her coauthors explore technologies used in policing, in prisons and to monitor people after release, with a focus on their effects on people.
UW News spoke with Eiger and Franziska Roesner, senior author on the papers and a UW professor in the Allen School, about the three papers and the themes that emerged across them.
“You can send the same message to 10 different people, and only one or two will get it, or one will get it 10 days from now, and five will get it rejected without knowing why. But… you pay to send each message — no matter if it drops or the call fails or the message is blocked. ”
Can you explain the studies?
Yael Eiger: The first paper focuses on technologies like Securus tablets. Securus basically owns the market for technology that incarcerated people use to communicate with their loved ones, lawyers or anyone in the outside world. But reports have shown that one in three families with an incarcerated family member goes into debt just from communicating with them. They’re charged just to send an email. So we interviewed formerly incarcerated people and loved ones of currently incarcerated people about their experiences with these technologies.
For the second, we ran workshops with people when they were released from prison. We started realizing how much digital literacy was playing a role in people’s outcomes after release. I would have students who’d maybe used a pager, then were incarcerated for decades and released into a society where artificial intelligence existed. So navigating things like applying for housing, jobs or food stamps was so difficult without baseline digital literacy. One of the things we also kept hearing about was people falling for scams often — phishing scams over text, email or phone.
For the third, I went to a big conference expo where companies try to sell their technology products to prison and police officials. I collected all the advertising literature, and we did a taxonomy of the technology being sold. We came up with a set of case studies of this technology, looking at some of the most pressing and urgent harms or risks of using it — whether that’s a First Amendment problem or a Fourth Amendment right-to-privacy problem in the legal setting, or bias and fairness risks of using AI.
Could you give me an example of a product at that expo?
YE: I was spamming Franzi the whole time with pictures I was taking. One that jumps out is an AI chatbot called AIDA. It’s basically a booth, maybe six feet tall and two to three feet wide, and you sit on a chair inside it, all encased in bulletproof glass. The premise is that when people are newly incarcerated or transferred to a new prison, they normally have to talk to a caseworker about their educational attainment goals and history. But this takes time, and often incarcerated people are distrustful of this process and don’t share much. So this company, Vantage, created a chatbot to have this conversation with them and elicit information they might not otherwise feel comfortable telling an officer, recording and analyzing it from a bunch of different sources.
That stood out to me because I was physically within this technology booth, and the privacy policies — how the AI processes, saves and retains the data, who it’s shared with, what AI model is being used — were not clear at all.
Most prisons are public but they’re filled with private products like those I saw at the expo. There are tons of public-private partnerships and contracts between private companies and public entities, paid for by taxpayers, but we don’t have much transparency into the technologies, the contracts, how data sharing happens, or whose rights are being respected, consumer protections and so on. It’s a very opaque ecosystem.
What are some important takeaways from these three papers?
Franziska Roesner: Often, when we think about consumer security and privacy, we’re thinking about technologies that consumers themselves buy and use. But in the carceral context, it’s very much technology being used by, or on, people who have no control or choice in using it. The people making decisions about buying it, and what features to prioritize and what failure cases matter, are just not aligned with the people it’s used on.
YE: A lot of times in security and privacy research, we’re worried about someone hacking into an insecure system. But in this case, the security and privacy problems are baked in, so there’s often a feeling of helplessness.
FR: Also, the incentives at play aren’t always toward improving society, improving people’s lives or determining the best strategies for rehabilitation. There are financial incentives and power dynamics at play.
YE: One of the clearest cases of that from our papers has been the usability problems with Securus. You can send the same message to 10 different people, and only one or two will get it, or one will get it 10 days from now, and five will get it rejected without knowing why — all these usability problems. But Securus doesn’t have much incentive to fix any of this, because you pay to send each message — no matter if it drops or the call fails or the message is blocked.
FR: But we know that incarcerated people being connected to their communities is actually good for their return to society later. So we, as a society, as tax payers, should want this technology to work well.
The research
- The first paper, presented at the Conference on Sociotechnical Cybersecurity and Privacy, explores security and privacy in technology available to people in prisons.
- The second paper, presented at the Twenty-Second Symposium on Usable Privacy and Security, looks at how technologies used on or by people after prison make them susceptible to scams.
- The third paper, presented at The 2026 ACM Conference on Fairness, Accountability, and Transparency, catalogues 122 products presented at the U.S.’s largest annual carceral technology expo.
Several of your co-authors were formerly incarcerated. How did that collaboration affect your research?
YE: It felt quite natural. These co-authors were instrumental to this research — whether through data analysis or, in the scams paper, organizing workshops and partnerships with community organizations. They contributed more to the paper than a typical co-author would. So it was clear to me that they were going to be authors, if they wanted that.
There’s a history of researchers going to work among marginalized groups and taking advantage of their expertise and their histories without crediting them, which in my opinion is both unethical and makes for worse research.
FR: There’s also always a risk when you’re working with marginalized populations that you come in, write the paper, get research credit for it in your academic community. And you get separated from what should be the actual goal, which is improving whatever it is you’re studying.
What do you want the public to know about this research?
YE: It’s always my goal to increase compassion and understanding about people we may not know or normally encounter. Having people speak for themselves, with a researcher able to lend a stamp of legitimacy or just listen to those experiences, is important to me.
Also, there are very tangible ways we can increase incarcerated people’s security and privacy in a compassionate way that doesn’t have to be at odds with the goals of lowering crime, preventing crime and making society safer. We have rights as people — free speech, a right to privacy. I’d like policymakers and technology providers to see the opportunity to make things better, and the general public to have a more compassionate understanding of these kinds of problems.
That’s not an unrealistic expectation. For instance, policymakers in 2022 released caps on the price companies can set for phone calls coming in and out of prison. So there’s a lot of legislation in this space that I think could make things a lot better.
FR: The general public’s exposure to these various surveillance technologies also raises concerns. We’ve seen that with Flock cameras and the upset that has been raised, and cities backing away from them. These technologies affect everyone.
A full list of co-authors and funding is included in the papers.
For more information, contact Eiger at yeiger@cs.washington.edu and Roesner at franzi@cs.washington.edu.