Crime Algorithm, Space Junk, and Stem Cell Therapies
Algorithms are used to predict the behaviors of certain groups of people, including shoppers, drivers, and office workers. In several states, parole boards employ this type of predictive data to determine “high risk” offenders who may commit another crime once they’re released. Maggie Koerth-Baker, a senior science reporter for FiveThirtyEight.com, discusses a case in the Wisconsin Supreme Court that challenges the legality of these risk algorithms. Plus, a look at the growing availability of stem cell treatments, for good or for ill.
Maggie Koerth is a science journalist based in Minneapolis, Minnesota.
Leigh Turner is an associate professor of bioethics at the University of Minnesota in Minneapolis, Minnesota.
IRA FLATOW: This is Science Friday. I’m Ira Flatow. Coming up, we’ll be talking about traumatic brain injuries in the military and PTSD. But first– algorithms are used to understand and predict their behaviors in all types of settings when you’re shopping at the mall or driving in traffic, sending emails at work. And in certain states, law enforcement is using this type of predictive data in sentencing criminals.
A court case in the Wisconsin Supreme Court is challenging the legality of these risk algorithms. Maggie Koerth-Baker is here to fill out that story. She’s a Senior Science Writer for fivethirtyeight.com. She joins us from the NPR studios. Welcome back.
MAGGIE KOERTH-BAKER: Hi, thanks for having me.
IRA FLATOW: You’re welcome, Maggie. So how are law enforcement and parole boards using this type of data?
MAGGIE KOERTH-BAKER: So these algorithms have been around for a really long time. Parole boards have used them for decades. And it’s basically these questionnaires that have all this information about an inmate’s background, criminal history, all this personal information. And they use it to come up with an estimate of how likely that person is to commit another crime.
So that’s one thing when you’re talking about parole and what you’re using it for is to kind of figure out what services a person is going to need. And it’s kind of another thing when you’re applying it to sentencing. And in the past decade or so, more and more states have started applying those things to sentencing.
And that becomes a problem because many of the factors that are correlated with the risk of committing a crime– things like the neighborhood you live in, the level of education you have, whether your family members have ever been convicted of crimes– those are also just things that correlate with being poor and being a person of color in America.
ProPublica had a big investigation of this back in May that everybody should really read. And they found examples where you have a white person and a black person being convicted of the same crime. And the white person actually had a more and worse criminal history. But the black person would get a longer sentence. And that is the kind of problem that we’re dealing with these algorithms.
The other thing is that how the algorithms work is often secret. So that’s the primary problem in this Wisconsin case– is that these are often proprietary systems that are sold to states by companies. And in those cases, you don’t know which questions are most important to the final outcome of the risk assessment.
So if you have a sentence that’s based on one, you can’t properly appeal it. Because you don’t know which questions mattered the most.
IRA FLATOW: So what will the Wisconsin State Supreme Court decide? What is the case about?
MAGGIE KOERTH-BAKER: So the case is really about whether you can use particularly secret proprietary algorithms in sentencing. But it’s also got some broader applications. I talked to a researcher named Christopher Slobogin at Vanderbilt University. And one of the things he was telling me is that this relates back to the Daubert standard for expert testimony, which comes out of the 1993 Supreme Court case.
And it basically reorients the standard of scientific evidence in the courtroom from what is generally accepted to what is scientifically valid. And the judge is the person in charge of determining scientific validity. So this Daubert standard is why you’ve seen so many previously accepted forensic sciences go up in flames in recent years. But it’s never been strongly applied the sentencing before.
And Slobogin says that this case is going to make a difference for how that Daubert standard gets applied to sentencing.
IRA FLATOW: Interesting. We’ll be following up on that. Let’s move on to up Pat Summitt, who was a legendary basketball coach at the University of Tennessee who passed away this week from early onset Alzheimer’s. Which is a rarer form of the disease, correct?
MAGGIE KOERTH-BAKER: Right. So there are about 5.4 million Americans who have Alzheimer’s. But about 200,000 of them are people who got Alzheimer’s before the age of 65. So Pat Summitt died at age 64. She was diagnosed at age 59.
So it’s this small subset of people with Alzheimer’s. But it’s actually a really important group as far as scientific research is concerned. And that’s because Alzheimer’s is really hard to study.
So you know how over the last few decades, scientists have decided with cancer that it’s not just one thing. It’s actually this whole array of different diseases that are kind of under one umbrella. Alzheimer’s is a single disease but it has lots of possible origins. It can look really different from one person to another. And there’s just this shared pathology that everybody has in the end the kills them.
So early onset Alzheimer’s is more likely to be genetically linked. And there’s actually sub-subset of people who have dominantly inherited Alzheimer’s. So if they have this gene variant, they’ll get Alzheimer’s. They’ll get it young. And there’s a 50% chance of passing it on to their children. And that’s a real personal tragedy.
But it’s a scientific opportunity because it allows scientists to start following the progress of the disease before people actually know and notice that they have it, so you can test somebody for this dominantly inherited Alzheimer’s when they’re in their twenties before they have any symptoms at all. And know that they’re going to develop it.
And so you can start watching for biomarkers that can help with testing. You can start figuring out possible treatments. You can start better understanding what happens in the brain before the symptoms or even obvious.
IRA FLATOW: The 5% helping out the 95%.
MAGGIE KOERTH-BAKER: Right.
IRA FLATOW: Let’s move on to China. China launched a spacecraft to clean up space junk. I mean, they have created a lot of space junk themselves, haven’t they? So what’s the plan?
MAGGIE KOERTH-BAKER: Everybody has created a lot.
IRA FLATOW: They’ve blown up satellites in space.
MAGGIE KOERTH-BAKER: Right. Yeah, that 2007 case where they decided to try to destroy space junk by blowing up a satellite that then created lots and lots of pieces of tiny space junk. Yeah.
Space is a huge problem. It’s all around in the lower third orbit. We’re talking about whole satellites that have been decommissioned and they’re just up there. But we’re also talking about a tiny little pieces of things from satellites crashing into each other. From wrenches that got lost. It’s not just big stuff. It’s also small stuff. And they can damage spacecraft. And they keep crashing into each other and creating smaller and smaller pieces that are still dangerous.
So you have this new satellite, which is called Roaming Dragon, and it has a mechanical arm on it. And it’s supposed to be able to pull up alongside a dead satellite and basically you pick it up like a garbage truck. And then Roaming Dragon could be piloted into a safe crash landing in the ocean. And it takes the junk with it. Which sounds great, right? Like that sounds like a really good idea.
IRA FLATOW: But.
MAGGIE KOERTH-BAKER: But.
IRA FLATOW: Any but to that?
MAGGIE KOERTH-BAKER: There is, in fact, a but to that. Last year I did the story on space law. And the space lawyers, which is a job, at the time, were telling me that the problem with any system that can clean up space junk is that it also inherently can operate as a weapon. That it can be something that could take out a working satellite just as easily as it could take out a dead one.
And that is a potential problem with Roaming Dragon. And it’s something that observers of space law and observers of what’s going on in space are kind of currently debating right now– whether or not this is something we should be worried about with this particular satellite.
IRA FLATOW: I’ve seen that James Bond movie, by the way. Where they actually did that in the movie.
Thank you, Maggie.
MAGGIE KOERTH-BAKER: Yeah, you’re welcome. Thanks for having me.
IRA FLATOW: You’re welcome. Maggie Koerth-Baker is a senior science writer for fivethirtyeight.com, based out of a Minneapolis.
Now it’s time to play a Good Thing, Bad Thing.
Because every story has a flip side. It’s a busy time for research into stem cell treatments for disease, with hundreds of clinical trials underway. Investigating their potential for multiple sclerosis, heart disease, Crohn’s disease– lots more.
Pro athletes from football to baseball are receiving high profile stem cell infusions and injections to treat injuries picked up on the field. Late hockey great, Gordy Howe, received them to treat a stroke. And a new report released yesterday has found 577 clinics– over almost 600 clinics, right here in the US– offering some form of stem cell treatment for a huge range of conditions.
Here to break down the good the bad of stem cells is my guest, Leigh Turner, an Associate Professor of Bioethics at the University of Minnesota. He is a co-author with UC Davis stem cell researcher, Paul Knoepfler, of a report, which appeared in the journal, Stem Cell This Week. Welcome to Science Friday.
LEIGHT TURNER: Hi, thanks very much for having me on.
IRA FLATOW: OK. Let’s start with the good. What are the known medical benefits of stem cells?
LEIGHT TURNER: You yeah, that’s a good place to start. I mean, most of us may know, from our own family experiences, that there are some stem cell transplants. Usually we think of those as bone marrow transplant. So if you think about family or friends, you may know someone who’s had leukemia, lymphoma, other blood-related disorders. And they may have had a bone marrow transplant. So that’s where the blood producing stem cells are transplanted. That’s part of evidence-based medicine.
There are also, in some instances, essentially bone marrow, umbilical cord blood containing blood stem cells that are transplanted. So again, you can think of this as sort of mainstream supported by safe and efficacy data.
IRA FLATOW: But that’s not what you found in the 570 clinics that they’re offering. They’re not really claim to do a lot of those things with stem cells.
MAGGIE KOERTH-BAKER: And that’s exactly right. I mean, we had an inclusion exclusion criteria. So we didn’t include procedures like that. We didn’t include state medical centers and hospitals doing those procedures. We focused on businesses making, really, quite dramatic claims. Like claiming they were advertising the stem cell therapy for ALS, Parkinson’s disease, multiple sclerosis, Alzheimer’s, dementia, cerebral palsy, autism, and the list goes on. 20, 30 different diseases and injuries.
And that’s where the evidence breaks down. These aren’t institutions that have done pre-clinical research or clinical studies.
IRA FLATOW: So that’s the bad news. So even though they’re making these claims, there’s really no good hard data to say that that stem cell usage will be productive.
MAGGIE KOERTH-BAKER: Yeah. I mean, it’s a complicated marketplace. And I wouldn’t say there’s sort of one size fits all. There are a lot of different businesses out there. The kinds of assertions they make vary. But if you look on the end I just mentioned, I mean, there’s just not meaningful safety and efficacy. And those are the ones that most concern us.
IRA FLATOW: Is there a way to change this to make this stuff work?
To test them better?
LEIGHT TURNER: We have mechanisms in terms of carefully designed studies. Properly conducted clinical trials where you start with pre-clinical research then you move to first in human safety studies. And move on, phase two, phase three studies. I mean, there are pathways that can be used. They contain risks. People can be harmed in any carefully conducting studies.
But you try and minimize those risks. And when you do all this outside of a more mainstream clinical research translational research framework, you run the risk of ratcheting up risk rather than trying to limit it every way you can.
IRA FLATOW: All right, Dr Turner. Thank you very much for taking the time to be with us today.
LEIGHT TURNER: Thank you, I appreciate it.
IRA FLATOW: Leigh Turner, Associate Professor of Bioethics at the University of Minnesota.