One of the first cases Valerie Newman looked into after she was appointed to head up the Conviction Integrity Unit in Wayne County, Michigan, was that of Aaron Salter. In 2003, Salter had been convicted of a murder in which two gunmen shot up a home in Detroit. He was convicted based on a dodgy witness identification, and despite having a solid alibi that would have been backed up by several witnesses had his trial attorney bothered to interview them.

As she began looking through the police file for Salter’s case, Newman found a folder labeled “Perp” — which is of course police lingo for “perpetrator.” There, sitting atop the stack of papers inside was a photo of the man police clearly believed to be the real killer. It was not Aaron Salter. Someone had attached a sticky note to the photo with the handwritten words, “Do not copy.”

For Newman, this was alarming on a number of levels. It was compelling evidence that Salter was innocent. It was compelling evidence that someone in the police department knew Salter hadn’t committed the crime and had deliberately concealed evidence of his innocence. Perhaps most alarming was what the file revealed about the justice system’s capacity to correct its mistakes: For the entirety of the 15 years Salter spent in prison, the clear evidence of his innocence was sitting at the top of a police file. It’s just that no one had bothered to look.

And then there’s the case of Terrance Calhoun. “Mr. Calhoun was convicted in two separate cases of sexual assaults of minors,” Newman told me. “In one case a condom was used, and the young lady reported it to the police immediately. The police reported to the scene and found the condom, right where she said everything had happened. It was sent out for testing.” But before the test results were returned, Calhoun’s attorney persuaded him to enter a no contest plea. “He had a very low IQ, around 67,” Newman says.

When the results did come back, no one checked the results. “I don't know where the test results went in the prosecutor's office,” Newman says. “Did it make it into somebody's file? We don't know. We were never able to determine if any prosecutor actually looked at the results.” Calhoun’s appellate attorney then filed an Anders brief, a filing in which an attorney tells the court they couldn’t find any issues to appeal.

Calhoun spent 15 years in prison before anyone looked at the results. The DNA results not only cleared Calhoun, they were a match to a serial rapist. “It was a very, very tragic case,” Newman says.

Even after his exoneration, Detroit police tried an explicable last-minute maneuver to keep Calhoun in prison:

In both cases, Newman’s office alerted courts to exonerating evidence that should have been discovered years ago ir police, prosecutors, or defense attorneys had demonstrated even the slightest bit of diligence.

One thing you discover pretty quickly when covering wrongful convictions is that for every exoneration, there are probably dozens of other innocent people who will never get relief. For each case in which an innocent person was convicted, the state likely put on a half dozen or more witnesses who testified to a jury about that person’s guilt — from medical examiners to practitioners of dubious forensics, to cops who seem especially “gifted” at eliciting confessions or finding jailhouse snitches who claim to have heard a confession.

Inevitably, you learn that the bad actors in the case you’re looking at also testified in dozens, hundreds, or even thousands of others. Sometimes you learn that an expert in a particularly shady field of forensics also trained and certified dozens or hundreds of other witnesses in that field, who then went on to testify in who knows how many other cases. It can quickly be overwhelming.

Not all or even most of those other cases are wrongful convictions. But there’s a good chance that quite a few of them are. And absent DNA — and sometimes even with it — most of those people will never get relief. The mere fact that an expert of cop or prosecutor misbehaved in one case — even egregiously misbehaved — typically isn’t enough to affect other cases involving those same bad actors. All the other people whose convictions were tainted would still need to overcome the significant procedural hurdles to getting their cases reopened. And most would need to do it without a lawyer.

On the defense side, the revelations brought about by DNA testing gave us the innocence movement — the legal clinics, Innocence Projects and pro bono work lawyers do to clear the wrongly convicted.

But the DNA revelations should also have prompted action on the government side. Every exoneration should have come with a thorough, comprehensive review of every state actor involved in that case, ideally funded by state legislatures or courts.

That hasn’t really happened. Instead, we’ve gotten Conviction Integrity Units (CIUs) like Newman’s. Typically housed within the offices of district attorneys, CIUs are charged with finding and correcting wrongful convictions and overly harsh sentences. The first such unit was established by Craig Watkins, the late defense attorney who successfully ran for DA in Dallas County, Texas. Watkins’s unit quickly made Dallas home to more exonerations than most entire states. The idea spread among other reformist DAs.

CIUs are a welcome concession from prosecutors that the system does sometime get it wrong. When they have the full backing of the elected DA, they’ve been remarkably successful. The approximately 700 exonerations they’ve won (according to the National Registry of Exonerations) have provoked some needed skepticism of the criminal legal system, and even occasionally inspired some reform.

But they’re also woefully inadequate. CIUs are tasked with reviewing decades of convictions, and most have waitlists thousands of cases deep. Most are forced into litigation triage, taking only the slam-dunk cases, or cases in which much of the investigative and legal legwork has already been done by pro-bono attorneys, law students, or nonprofit firms.

Worse, over the last few years, even these low-budget, comparatively tiny offices have begun to face backlash. As I reported earlier this year, in one of his first official acts upon taking office, Virginia Republican Attorney General Jason Miyares fired the entire staff of the CIU set up by his predecessor, then put the office under the authority of a law-and-order prosecutor from Arlington who had recently been defeated by a reformer (and served on Donald Trump’s commission for police reform).

CIU staff I’ve spoken to in states like Texas and Florida say the aggressively law-and-order AGs and governors in those states have intimidated their elected DAs, prompting the CIUs to keep a low profile.

But CIUs and DAs are also getting pushback even in more progressive jurisdictions. As exonerations reveal corruption and incompetence, police departments are getting less cooperative. Budget-minded city officials have grown more skeptical of CIUs as exonerations bring multi-million dollar settlements and jury awards. Judges accustomed to our adversarial system can get defensive, or perhaps are just ill-equipped to handle cases in which both prosecutors and defense attorneys agree that the system has failed.

To discuss all of this, I spoke with Marissa Boyers Bluestine, Assistant Director for the Fair Administration of Justice at the Quattrone Center at the University of Pennsylvania Law School. In that position, Bluestine studies, tracks and works in consultation with CIUs around the country.

(Full disclosure: I’m a journalism fellow with the Quattrone Center.)

I’ve edited our Q&A for clarity and length.

How many CIUs are there now? Are they growing or shrinking in number?

By our count, there are 128 now. A couple of those identify more as a process than a separate unit, but we still count those. There was a big rush of units between around 2008 and 2010, then it was relatively steady for a while. There was also some rapid growth from about 2018 through 2022. There have been a few new ones this year, too — one in Hennepin County, Minnesota; a couple in Massachusetts; and one in Ohio, Lucas County, just announced this year.

So I don’t know that I would say it’s slowing down. It’s been pretty steady growth.

Before we go further, I’ve heard these called both CIUs — Conviction Integrity Unit — and CRUs — Conviction Review Unit. Which is correct?

We prefer CRU, but I think we’ve probably lost that battle. They’re interchangeable.

I’m sure it varies from unit to unit, but on average, how many exonerations does a typical CIU get per year?

It’s probably five or fewer per year. But I actually push back pretty hard on the notion that exonerations should be a metric of success for these units. An exoneration is such a luck event. Every domino has to fall in exactly the right manner at exactly the right time. It’s much more common that a CIU will get to a point where they lose faith in the conviction, but there’s enough residual evidence that they’re not comfortable giving the guy a full walk, either. So it might be a withdrawal and vacation of the current conviction and sentence, and then a re-plead on a lesser charge, usually for time served or a reduced sentence on an Alford plea. We call these “case corrections,” and they’re much more common than exonerations.

We’re trying to track those, but it’s pretty difficult. We do a survey every couple years and we try to get people to self-report, but it’s hard. But I just had a conversation with Val Newman in Detroit. She says that they’ve had 30-some exonerations, which is a lot! But they’ve done even more case corrections. It’s a similar story in Boston, and with several of the units in Florida. Even Philadelphia, which is touted as one of the most successful CIUs right now, a lot of their exonerations don’t appear on the National Registry of Exonerations because they’re more like case corrections.

So to answer your question, I’d say the typical CIU gets five or fewer exonerations per year, and maybe half to three quarters of units don’t have any most years. But that doesn’t mean they’re not doing important work.

So just a rough estimate, how many CIU case corrections would you say there are for every exoneration?

Three to four would be my guess. But again, it’s hard thing to say for certain.

Because CIUs generate headlines when they win exonerations, I suspect many people think they’re much larger than they are. What does a typical CIU look like in terms of personnel?

They’re tiny! It’s one, maybe two attorneys and usually an investigator that is either borrowed from another unit in the office or works under contract. Maybe there’s a part-time admin or paralegal. But the vast majority of these units are very small.

So it’s one or two attorneys, an investigator, and maybe an admin, and their job is sort through decades of cases for potential wrongful convictions?

That’s right.

My impression is that many CIUs work in coordination with Innocence Projects and clinics. Do they contract work to outside attorneys?

It would be a conflict to contract with innocence groups, but they do collaborate on cases. So an innocence group will bring a case to them, and then they’ll work on it collaboratively. That’s becoming the norm. In Minnesota, the State Attorney General’s office has a group of volunteer attorneys who step in to help out. And some CIUs are using students from legal clinics run out of law schools. Wayne County, other Michigan units, and the Philadelphia unit use interns. But the vast majority of CIUs are still just very small units in small offices.

That seems woefully insufficient given the enormity of what they’re asked to do. Even the most conservative estimates of wrongful conviction rates would still mean thousands and thousands of cases.

Yes. Right now there are 128 units out of around 2,700 prosecutorial agencies in the United States. That’s not even 5 percent. So that number needs to grow. But you also have DA offices that are so small — they have maybe four or five attorneys in the entire office. So creating a separate CIU just isn’t possible. They don’t have the resources, but also, it just isn’t possible for an office that small to review a conviction objectively because everyone in the office knows about every case.

So we need to look at different models. We’ve looked at regional models, for example, where offices collaborate together — County A reviews County B’s cases, and back and forth. But there are no working models of that just yet.

There’s also the whole question of the fox guarding the henhouse. Is it really even possible for a prosecutor’s office to undertake a project like this? I think the answer is yes, but only if they put in the resources and time to build a proper unit.

There’s also an argument to be made that this should be more of a judicial function — where you’re looking for facts and not letting pesky rules of evidence keep out things that otherwise might be probative. So far, the only unit under that model is the state commission in North Carolina. They took a look at cases as a whole, not in the sense of what they’d need to file a petition, and it’s all funded by the legislature, not prosecutors. So it’s an entirely different way of looking at it.

A state CIU with jurisdiction to review cases statewide does seem like a good way to cover smaller DA’s offices who don’t have the resources for a CIU. How many statewide CIUs are there?

Let’s see — there’s California, Illinois, Michigan, Minnesota, Massachusetts, New Jersey, and Pennsylvania. Connecticut and Delaware also have a statewide CIU, though in those states that’s really the only option because they have a unified prosecutorial system. The New York Attorney General has an office, but I haven’t had any contact with them. And then there’s Virginia’s “CIU” which I’ll put in air quotes, because it was “restructured.”

Most of these state units have come online pretty recently, so I think we can say there’s been a lot more recent expansion on the state side than on the county side.

Do these state offices generally avoid jurisdictions that already have a CIU and try to focus more on counties are districts that don’t?

Yeah, that’s exactly right. So in Pennsylvania, for example, there are four county-level CIUs in the state. So if a request comes to the state CIU from one of those counties, they’ll refer the request back to the county. They really see themselves primarily as a resource for counties that are too small to have a unit of their own.

Are there any CIUs in the federal system?

There are a few federal offices. There’s one in New York, one in California, and one in D.C. Under the federal system you have a higher evidentiary standard and other safeguards that you don’t have in many states, but mistakes still get made. So these units are still important.

One place where there is no CIU and absolutely needs to be one is in the military justice system. There are no conviction integrity units and no innocence organizations in the military system at all. It’s entirely different from the civilian system, so innocence organizations won’t even look at military convictions. But I’ve heard about some really some egregious cases that absolutely need to be reviewed. But right now there’s no mechanism to do it.

That’s interesting. The military system gets very little coverage from journalists, either. Probably for the same reason. Most of us just don’t know much about it.

Do any of the federal CIUs review state convictions?

No, they only review federal cases.

CIU directors I’ve spoken to have said that while they get applications from both defense attorneys and innocence groups, they get a lot more “pro se” requests — applications from prisoners who don’t have a lawyer. But because the CIUs have limited resources, they have to prioritize. And there are a lot of competing arguments about which pile should get more attention.

They probably tell you they don’t take action on the people who write individually because they just don’t have the time. They focus on cases brought to them by lawyers, or even from within the office. Some have been able to push through some exonerations from pro se applicants, but most of the people they exonerate are represented.

Right. There’s a push and pull at work. The requests from people who already have a lawyer are pre-vetted and often much of the legal work is already done. So they’re easier “wins.” But these are also people who already have help. The pro-se applications may be more work and take more time to vet, but those are people who would otherwise never be heard. It seems like there’s an argument that they should be prioritized.

Yeah. I understand that. But I think it’s just the reality of the situation. When we started this project, we saw there were units who weren’t even opening the letters that were coming in from pro se applicants. They just didn’t have the staff for it.

So we created guidelines for how to work with unrepresented and underrepresented individuals. The latter is almost equally a problem, by the way. You also can be represented by an attorney who doesn’t give your case the attention it deserves, and that can be just as bad.

But it isn’t just about resources. The CIU is still part of the DA’s office. It’s still part of the prosecution. So if you’re reviewing a case from a pro se prisoner, you have to do so in a way that is ethical and respects their rights. When a CIU staffer finds a pro se case that they want to take on, they can’t move forward ethically until the prisoner has his own representation. Someone has to be looking at the case from his perspective.

But at the same time it’s also problematic for a CIU to then directly refer that case to a specific defense attorney.

That’s a difficult conundrum. You can’t move forward on a pro se case until the prisoner has an attorney, but you can’t refer them to a specific attorney, either.

Right. But you can give them a list of people who have agreed to take those cases. So the model we’ve recommended is to have a list of attorneys who have offered to represent people pro bono, along with the name of a local innocence organization if there is one. So when a CIU finds a pro se case that has merit, they’ll write back to the prisoner and send them the list. It’s not a perfect model, but it seems to work.

Some of these units have waitlists that are hundreds or thousands of cases long. How else do they sift through the applications. Do they prioritize some cases over others? Or maybe the better question is how should they prioritize some over others?

Yes, those are two different questions.

Sometimes a particular CIU will prioritize cases based on the expertise of staff and what the well-known problems in their jurisdiction — maybe there’s a problematic forensic specialist who has testified often in local courts, or maybe there’s a police detective who has elicited false confessions in the past. We’ve had big crime lab scandals in Massachusetts and Colorado, so in those states those cases are going to get immediate priority.

Most CIUs have a long form for applicants or their lawyers to complete. It tends to rely heavily on narrative. So someone in the unit will go through the entire narrative and then decide whether it makes sense to follow up. At this stage, folks tend to rely on what we call canonical factors — eyewitnesses, false confessions, bad forensics — the stuff you see over and over in these cases.

But that isn’t always efficient. Let’s say you’re reading through a narrative and you see that this an eyewitness ID case. That’s a really unreliable type of evidence, so you keep reading. A few hours later, you’re three quarters of the way through a banker’s box of files and you learn that the eyewitness was the suspect’s sister. That would have been good to know much earlier.

So the question is, could CIU staff use their limited time in better ways? Are there shortcuts we can use to quickly isolate the most promising cases?

One way to do this is to have applicants check off boxes. There are specific questions that can give you a sense of the relative value of diving deep into a case. So we’re trying to figure out the best way to do that. We’re trying to see if we can come up with an algorithm or an evaluation tool that can look at these cases objectively and figure out which are most likely to be wrongful convictions.

We know, for example, that eyewitness identifications are common in wrongful convictions. So we’re working with experts like [former police chief] Bill Brooks in Massachusetts to create a robust checklist. What kind of eyewitness identification was it? Was it cross-racial? Was the witness given a lineup? Was it a lineup with photos, or was it a show-up?

We also know that false confessions are common in wrongful convictions. So if they check that box, you get additional questions — was the suspect a juvenile when he confessed? Was the suspect mentally ill or disabled? Was it a long interrogation? Was the questioning abusive or coercive?

The tricky thing is trying to figure out which common characteristics of a wrongful convictions co-relate. How should we weight a case with a confession versus an eyewitness ID versus a case with both? Do false confessions co-relate to prosecutorial misconduct — or to something else?

The goal here is to get to the point of having a data-driven decision tree. So okay, this case has a false confession, so now we’ll look to see if there’s any prosecutorial misconduct. Now we’ll look at forensics. There’s a team at Stony Brook University working on this question.

Where should a CIU fit within the organizational structure of a DA’s office? There’s an argument that they should be independent to the point where they’re actually physically isolated in their own office to prevent traditional prosecutors from influencing them. But there’s a counterargument that separation just breeds resentment, and that it’s better to be integrated with the rest of the office.

These are all important questions. In his 2016 report for the Quattrone Center, John Hollway listed three main priorities: independence, flexibility, and transparency.

I think we’ve since added a fourth, which is protection against bias. It’s just critical that the unit not be influenced by confirmatory bias when they’re reviewing these cases. Most of these units are properly situated outside of the DA’s appellate unit, outside of the trial unit, and they independently report directly to the elected [the DA]. That’s the way it should be.

If you put your CIU in the trial unit, you’re at risk of falling into a prosecutorial mindset. “We don’t prosecute people who are innocent, therefore we couldn’t have gotten it that wrong.”

And if you put your CIU in the habeas or appeals unit, their fallback position is to protect the conviction. “We’re here to protect the conviction, not challenge it.”

It’s just not kind of in their DNA to take a different look at these cases.

So it should be fully independent. The unit should also report directly to the elected. Why? Because if you have to go through the habeas unit to decide whether or not to take action, it’s still going to be looked at through that lens of protecting the conviction instead of potentially challenging it. Same thing with the trial division.

The problem is that in large jurisdictions like Philadelphia or Manhattan, it’s impractical for the elected to weigh in on every case as quickly as you might want. That can slow things down. So sometimes a deputy can take that role. But as long as they have the absolute authority from the elected to make that decision, and the elected chooses someone who isn’t going to be looking at these cases through the lens of the institution, that’s a possible workaround.

In many states, once a case gets to post-conviction, it’s handled by the state attorney general’s offices. But in states that don’t, or for cases still in direct appeal, what happens when a DA’s CIU wants to take on a conviction that the same DA’s office is still defending?

This is why it’s important that the ultimate call comes from the elected. It needs to come from a unified voice, from a unified vision that takes the whole office into account. What we often see, and what we think is actually a really good practice, is to have an intermediary advisory board.

So the unit says, “Okay, advisory board, this is what we want to do. Here’s the action we want to take.” And then the board reviews the case and says whether or not they agree. If they agree, the case moves on to the elected. That might help protect against any bias the CIU may itself have about finding exonerations, and it gives political cover to the elected. If it’s a particularly dicey case that they want to move forward on, then the elected pull back and say, “Well, wait a minute. This is what my advisory board wanted me to do, and I’m going to go with that.” It provides some cover.

Of course, things can always go sideways. There was a case in Ohio where the entire advisory board quit. They all resigned because they were making recommendations and the DA wasn’t acting on them.

Ultimately it’s the DA who’s going to make the decision. The question is who’s presenting the information to them, and has it gone through a second level of potential bias check before getting there?

Getting back to “case corrections” versus exonerations — this seems like another area where limited resources may require a CIU to prioritize one class of cases over the other. Exonerations make news and can inspire reform. But as you say, because our system puts a premium on finality a lot of things have to fall into place for an exoneration to happen. It’s quite a bit easier to persuade a sympathetic DA to go along with a sentence reduction. There are CIUs that focus almost entirely on reducing sentences, and others that focus mostly on exonerations.

What’s the right balance?

That’s such a hard question. We don’t recommend that you have kind of a one size fits all kind of unit. The worry is that if it is a particularly difficult case, if it’s politically touchy or it’s going to require a lot of resources, it’s a lot easier to adjust the sentence, even though a full exoneration would be more just. You don’t want to make that a fallback position.

But also, the criteria are very different when you’re looking at modifying a sentence. Guilt or innocence isn’t necessarily a part of that. It’s more about whether they were treated fairly. Was this a sentence that we would give today? How has this person adjusted in prison? Do they have a potential home plan? Do they have support in the community? These are very different factors.

But sentence reduction is a critical part of the prosecutorial function, especially in a second look jurisdiction where you have the ability to go back and reevaluate. In some places, like in Brooklyn, they have one section that only does sentencing reviews and another that just does exoneration work.

That may be the right model, because it’s important to look at these cases through the right lens. Maybe the exoneration section doesn’t think there’s enough evidence of innocence, but they do think this person was over-sentenced, so they can move the case to the other section. We’re be more worried about the fallback problem when it’s all done by the same staff.

Also, when the same people are doing both types of cases, you dilute the resources for the very hard work of innocence cases. It makes it easier to just focus on the sentence revisions that are more likely to get results. And that’s not good, because if you only have one or two people doing it, they just can’t do both.

In some DA’s offices, the CIU can’t take on any case that’s still in litigation. I find that hard to understand. If the state realizes that an innocent person is in prison, they should immediately go to court to get that person at out, not continue to defend the conviction just because the innocent person is in mid-appeal. Resolving appeals can take years. And it seems especially odd in death penalty cases, which are typically litigated right up to the execution.

I’m not sure if it’s that they they can’t take on those cases so much as that they won’t. I think it's probably won't. But it’s similar to the policy you see in innocence organizations. Most will not take on a case if there’s an active appeal because the theory is that, well, the appeal could work. There are reasons for that, mostly related to resource allocation. If this person can get relief through an appeal or habeas petition, then we can use our resources on someone else.

Yeah, that seems like an understandable and necessary policy for innocence groups.

On the prosecutor side, though, the argument is very different. You’d think they would want to find these cases. If you can identify an innocence case earlier in the process, you save years of effort and resources of addressing it now as opposed to later.

In New Orleans, where when Emily Maw took over [the CIU], they have now that policy. If a defendant raises an innocence claim on appeal or post-conviction — or a Brady claim or a racial discrimination claim — prosecutors now have to provide substantive reasons why they’re going to continue to defend that conviction. They have to investigate it. They have to determine if it’s a valid claim on the merits. They can’t just argue that the defendant should lose because he missed a deadline or some other procedural reason. I think that’s a good policy.

“They just threw everything out. They didn’t even keep track of which cases they had to throw out. It was easily over a thousand cases, probably multiple thousands of cases where they just incinerated all the biological evidence.”

I think Andrew Warren was also looking at that policy in Tampa before Ron DeSantis removed him. We’ve also talked to other offices about it. When someone makes an innocence claim or a Brady claim in post-conviction or on appeal, instead of reflexively trying to get it dismissed on procedural issues, the prosecutors will look to the CIU and say hey, what do you guys think of the merits of this claim? Should we send this case to you?

I would not say this is the policy in a majority of offices at all, but there are more and more offices looking at it, because it’s better for the prosecutor, the wrongly convicted, the community — everyone — if they can identify a wrongful conviction early on. It also gives them more time to find the real perpetrator. If you wait until the litigation is over, the evidence is old. You don’t know if witnesses are still around, or even still alive.

You’d think that part — catching the real perpetrator — would get more attention.

It’s encouraging that you’re seeing more of that sort of cooperation between CIUs and other parts of the DA’s office, even if it’s still only a handful of offices. But more broadly, based on what you’ve seen heard, how are CIUs and their staff fitting in with other prosecutors? Is there tension? I could see traditional prosecutors being resentful of this unit that’s looking over their shoulders.

There’s definitely tension. The head of one New York CIU says he’s referred to as “the scariest missed call in the office.”

That’s funny. I guess that at least means they respect him.

Right. They do respect him. When these units get started, there’s a lot of skepticism from the line prosecutors, no question about it. What we have found that really helps, though, is for the electeds themselves to stand with the CIU director — literally stand next to them at an all-staff meeting — and say look, this is how we’re going to do things now. If you get get a call from the CIU, you’re going to cooperate. You will not come to me to try to do an end around. You will respect them and their process. The more that the elected sets the tone, the less you’re going to see pushback.

But yes, we do see it all the time. The CIU is investigating a case, and the trial attorney or appellate attorney doesn’t think they should be. So they go to the elected and try to do an end run around the investigation. It’s then up to the elected what to do about it. If they believe in the process and they believe in the unit, they’ll say, “Nope. You have to respect it. You’ve got to go through this.”

When they allow that kind of whispering in the ear, that's when tensions rise and can eventually boil over.

It’s one thing for a reformist prosecutor to set up a CIU. But do these offices have staying power? What happens to these units when a more conventional prosecutor gets elected? Does the CIU get disbanded? Marginalized?

Sadly, we have seen some offices marginalized, like in Boston. When Rachael Rollins left, a new DA came in and demoted the person Rollins had brought in to run the unit. And I don’t believe the unit is as well staffed as before. When a new DA took over in Suffolk County, New York, they didn't disband the unit. The unit is still there and operating, but there has been a less than full-throated endorsement of it. The only one I've seen actually disbanded was of course in Virginia. And I guess that one wasn’t technically disbanded, but the new attorney general fired everyone, renamed it, and put it under someone with lots of unrelated responsibilities.

It will be interesting to see what happens in Cook County, Illinois. The CIU there has been involved in so many exonerations. If the new DA pulls back, it’ll be interesting to see how the public reacts.

How often have you seen an elected DA overrule the CIU when it wanted to pursue a possible wrongful conviction?

It has definitely happened. It happened in Ohio.

That was a Shaken Baby Syndrome case, right?

Right. That’s the one where the advisory board resigned. But I can’t think of many other examples. The elected might slow-walk a case — or if it’s politically tricky they might try to kick it to the next administration to deal with. But I don’t know of any others where the DA has just said no.

Another source of pushback has been conservative state attorneys general. In many states, appeals or post-conviction or both are handled by the state AG, not the office of the DA that won the conviction. And we have definitely seen cases where the AG’s office ends up opposing the local CIU in court.

Oh yes. That’s exactly what we’re seeing in Missouri right now. We’ve also seen it in Utah.

Louisiana, Alabama, and North Carolina, too.

That’s part of the reason why we don’t like state legislators getting involved this, because when they pass legislation on wrongful convictions, they like to include that carve-out — they give the Attorney General the right to come in and intervene and defend a bad conviction.

We’d rather let district attorney offices set these units up on their own, as part of the prosecutorial obligation to seek justice. When you bring the attorney general in, you risk letting someone like the AG in Missouri intervene and oppose every exoneration regardless of merit, like with Marcellus Williams.

So yes, there is definitely tension in some of these states between the attorney general and the CIUs and the prosecutors’ offices.

I want to talk about some of the other challenges facing CIUs. Pam Colloff recently wrote a long piece for the NYT Magazine about a case here in Nashville, where the CIU, the DA’s office, and the defense all argued in favor of throwing out a wrongful conviction, but the judge refused. Does that happen often?

It definitely happens. There’s a case currently pending before the Supreme Court of Pennsylvania in which there was an agreement on the proper outcome between a conviction integrity unit and the DA’s office. Not an innocence case, but they agreed there was a Brady violation, so the DA is ready to retry the guy. But the victim’s family filed what’s called a King’s Bench petition in the Pennsylvania Supreme Court. They asked the court to assume jurisdiction and address the question of whether a trial court can proceed where there’s that sort of agreement between the parties. I think how that turns out could have a huge impact around the country. If they say oh yeah, the attorney general can intervene, or the victim’s family can intervene, that’s going to fundamentally change the law. We don’t have intervention in criminal law. It happens in civil court all the time, but on the criminal side, it’s the state versus the defendant. That’s it. There are no other parties. Victims may have a right to participate as amicus, or be heard in other ways, but they can’t directly intervene.

It doesn’t involve a CIU, but there’s something similar happening with the Richard Glossip case out of Oklahoma. The state attorney general think his conviction should be overturned due to overwhelming doubt about his guilt. But a state appeals court refused to go along. The Supreme Court agreed to hear the case, but there’s no one from the state to argue it. It’s just the courts. So the Supreme Court hired an attorney in California — a former clerk for John Roberts — who previously had nothing to do with the case to come and argue the position that this guy should be executed.

That’s correct. It’s an interesting question, right? Civil litigants can often settle without the okay from a court. But on the criminal side, you can’t do anything without judge’s permission. Once the charges are filed, that’s it. A judge has to sign off on everything — any agreement between the state and the defense. Charges being withdrawn, a plea bargain, it all has to go through a judge. So even though post-conviction is technically civil, it obviously has a criminal element, so a judge still has to be involved. And you can’t expect the judge to just rubber stamp every agreement between the prosecutor and the defense. But in most of these CIU cases, both sides have done their due diligence. They’ve certainly done far more investigation than the judge has. They’ve usually found far more relevant evidence than what came out at trial. So while I think it’s important an important function for judges to sign off on these agreements, you also hope that the judges will give a lot of deference to the work both the state and defense have done.

It seems like some judges have started to mistrust CIUs — to dismiss them as activist or too ideological.

I think some judges may are too invested in the adversarial process — the idea that every hearing should have two opposing sides. Some prosecutor are now saying that as officers of the court, they cannot allow this conviction to stand. They have an ethical obligation to take action on it. Some judges aren’t used to that.

I should say that most of the time judges go along. But you do get the occasional case, like in Tennessee, where the judge says, well, there was no adversarial process here.

But when there is no “other side” — if all the evidence of this person’s guilt has been disproven, it seems performative to have an adversarial process just for the sake of having one.

Right. And this is one of the myths we tell ourselves about the criminal justice system — that the adversarial process gets us to the truth. It’s not about getting to the truth. It never has been. It's about competing stories. If we wanted truth, we’d go into the inquisitorial system they use in Europe, which is much more geared toward finding truth.

So there is already this misunderstanding of the purpose of the adversarial process, and judges want to force these CIU cases into that process when it may not be appropriate. There was a CIU case in Pennsylvania in which the judge who granted the relief said, “Look, this is all new to me. I don't know what to do.” That’s on the record! And the irony is that with a thorough post-conviction investigation, we often get closer to the truth than before. Certainly that’s the case with DNA evidence.

So I think there is some confusion in the judiciary about what these CIUs are doing, and what their role is as a judge. They misunderstand and think they're being asked to rubber stamp, as opposed to do their own diligence and do the case review and determine whether the agreement makes sense, whether it's appropriate under the law, and whether all of those procedural checkmarks have been made.

So there has been some rebellion. I would say it’s very uncommon, but it’s concerning when it does happen.

CIU directors have told me they’re now getting pushback from police departments, too. So many of these cases involve bad or even corrupt police work, and police leadership can be reluctant to cooperate because they don’t want those details to sully the department’s reputation with the public — even in cases from decades ago. Some CIUs are also getting increasing resistance from city officials, even in very progressive cities, because officials worried about the civil liability and lawsuits that arise from exonerations. Both are understandable reactions. But of course neither of those concerns justify keeping an innocent person in prison. How are CIUs dealing with those pressures?

I think the position for the units is pretty clear — none of that matters. They have an ethical obligation to do what they're doing, regardless of who might be upset by their work. If I undertake this investigation and I find evidence that this person is innocent, number one, or even that the conviction is improper because of the gross violations of the constitution that occurred at the trial, I can’t just turn away from that because it would have political ramifications. I’d be violating my prosecutorial oath, and every prosecutor’s ethical responsibility.

But the reality is that they’re also in a difficult situation. You start racking up exonerations, the city starts having to pay, and now you start getting pressure to scale back the CIU. I get that, 100 percent. And you can’t stop compensating the wrongly convicted. They deserve to be compensated for the years that were taken from them. Maybe the answer is that you kick compensation to the state, because the state can absorb those costs.

You can also pass compensation laws that are robust enough that people aren't pursuing civil rights lawsuits because they’re either not getting their statutory compensation fast enough or it’s not enough to cover their expenses. So overall, I would say that one answer is more robust compensation law from the state.

One other interesting issue we’re seeing is that in jurisdictions like Pima County, Arizona, and some counties in Nevada, is that the prosecutor also represents the city. So the DA who oversees the CIU, whose work can lead to exonerations — and then lawsuits and compensation — would also be representing the city in those lawsuits.

So you get these scenarios where it’s like, we can’t turn this case over because if we do my other client, the city, will be harmed because they’ll be sued.

Wow. That’s a huge conflict, no?

Well it shouldn’t be, though. It’s apples and oranges. You’re putting decisions in the same basket that don’t belong in the same basket. Of course, you have to turn over the information. Maybe you give a heads up to the civil side. Maybe you say, hey, this is coming down the pike for you. But that doesn’t mean you can withhold evidence in the criminal case, or fail to produce it in the first place.

So we keep seeing these weird little ethical quandaries that pop up, all of which are being resolved on the side. But you still take those cases, because an innocent person shouldn’t have to suffer for them. So you tell the city we’re going to do our job ethically. We’re going to do our due diligence. We’re going to provide the information we’ve found, and we’ll have to deal with the fallout.

But you also say, we’re also going to learn from these cases. Exonerations are about justice in those individual cases, but they should also be something to learn from. So you also tell the city, we’re doing to learn from what went wrong. We’re going to adopt best practices so that we can prevent wrongful convictions from happening in the first place. We’re going to use best practices for witness identification procedures. We’re going to implement protections against false confessions. We’re going to use science-tested forensics in a robust, transparent way. We’re going to guard against prosecutorial misconduct with open file discovery. You want these cases to move police and prosecutors into a preventative mode as opposed to just a full reactionary mode.

Right — so you say doing the right thing in these cases now will prevent wrongful convictions — and settlements — in the future.

Some states have passed laws stating that in order for an exonoree to receive statutory compensation, they must agree not to bring a civil rights lawsuit. That seems problematic, not least because it’s often those lawsuits that expose the police or prosecutor misconduct that contributed to the wrongful conviction. And we’ve seen cases where that exposure has led to broader investigations of those police officers or prosecutors, which has resulted in more exonerations.

Yes — and now you’re subjecting the wrongly convicted person to this choice between compensation and a lawsuit right as they’re released, when they have no money and they’re just trying to survive.

I was involved in a case in Pennsylvania where the plaintiff asked the police department to institute some reforms with respect to how they conduct investigations, so they wouldn’t make the same mistakes again. The police department said no, we don’t want that reform. I think they just increased the settlement value instead.

In reporting on these cases over the years, I’ve consistently found when you find a case in which the actions or testimony of a cop, prosecutor, forensic witness — whoever — were responsible for a wrongful conviction, it’s almost never the only time that person has tainted a case. There are always other cases. How often does a wrongful conviction exposed by a CIU lead to a broader investigation of the bad actors who made that conviction happen?

I can’t give you any numbers, but just off the top of my head I can think of five to ten jurisdictions where that has happened. In Suffolk County, New York, for example, they found a wrongful conviction due mainly to misconduct by the prosecutor. That triggered an investigation of all his cases which found that he had withheld exculpatory evidence in 100% of the cases that he did. Every single case.

That’s incredible.

Now, not every example was sufficient to overturn a conviction. And of course not all of those were innocence cases. But he had withheld exculpatory information in every single case he took.

In Austin, Texas, they they actually got a federal grant to work with the Innocence Project of Texas to look at a whole group of cases in which the same group of police officers were consistently eliciting false confessions. And of course we’ve seen it in forensics — most famously in the Massachusetts crime lab with Annie Dookhan, but also in Colorado and Minnesota.

I think the main reason we don’t see more of that kind of broader investigation is that in many of these jurisdictions, the data just don’t exist. Even if you can identify a bad actor — let’s say Officer Bluestine was responsible for this wrongful conviction —you can’t just go back and review all the other cases Officer Bluestine may have been involved with because there’s no way to do that. In Philadelphia, they wanted to go back and review all the cases that involved cops known to be bad actors. But they couldn’t, because there’s no database where you can just type in “Bluestine” and get a list of cases. You can’t even do that for cases that went to trial. And of course most cases never go to trial. So if there was a plea, there’s not even going to be a trial record.

And innocent people plead guilty way more than we’d like to admit, for all sorts of reasons.

Right. What they would have to do is go into the file room and pull every homicide from, say, 1984 to 1992, and then look through to see if Bluestine appears in the record. They just don’t have the resources for something like that.

So instead they rely on anecdotes. They put word out with defense attorneys, “If you ever had a case with this cop, let us know. If you represented somebody who had this cop, let us know.” And then they can backtrack. But they don't have the ability to do the sort of full audit that we would like them to do. They just can’t.

It takes a lot of people being wrong to get a wrongful conviction — police, prosecutors, expert witnesses. When I’ve written about these cases, I inevitably find out that those people probably worked hundreds or thousands of other cases. It just seems woefully insufficient that after all these exonerations, the only response from the state has been these tiny, under-funded units that exist in less than 5 percent of the country. It can be overwhelming to think of all the innocent people sitting in a prison cell whose cases will never be reviewed.

It’s horrifying. In Philadelphia, when I first started with the Pennsylvania Innocence Project, there were cases covering several years in which the rape kits and other biological evidence had been stored at a unit on Erie Avenue. The unit got infected with maggots. So they just threw everything out. They didn’t even keep track of which cases they had to throw out. It was easily over a thousand cases, probably multiple thousands of cases where they just incinerated all the biological evidence. Nobody kept a record of which cases had biological evidence that was destroyed and which cases didn’t, so we’ll never know. Who knows how many innocent people’s cases were included in that unit?

In the Annie Dookhan scandal, the Houston crime lab scandals, the Fred Zain scandal in West Virginia and some others, you did at least have a thorough audit and investigation. And I understand what you’re saying that in some jurisdictions, they just don’t have the data to do an audit.

But in others it’s just been a matter of political will. There are the Steven Hayne and Michael West cases in Mississippi that Tucker Carrington and I wrote about in our book. There are the revelations about Mary Jane Burton at the crime lab in Virginia.

The reaction from state officials is, “Well, we don’t want to open Pandora’s Box.” They’ve literally said that. But by “Pandora’s Box,” they mean “people who were wrongly convicted.” They mean innocent people still in prison. They’re saying they just don’t want to know about them. I find that unfathomable.

Right. They don’t want to know how deep the problem is. I don’t know how you respond to that. I just can’t wrap my head around it. In the years I’ve been doing this, I’ve never been able to wrap my head around that.

Many of these units were set up by progressive prosecutors. And unfortunately, part of the recent backlash against those prosecutors and criminal justice reform has included a backlash against CIUs. We saw it in Virginia, when Attorney General Jason Miyares fired the entire CIU office. So I’m curious about states like Texas and Florida, where governors and AGs have taken a really confrontational posture against progressive prosecutors. Has that affected how the CIUs in those states operate?

There’s no question that it’s had an impact in Florida. DeSantis has removed two elected prosecutors there, and I don’t think it’s a coincidence that both of those prosecutors oversaw robust and effective CIUs that were actively working cases and reversing convictions.

I would say though that we’ve seen the opposite in Missouri. Even though the state AG is pushing back in these exoneration cases, you’re still seeing prosecutors and CIUs push ahead.

I’d guess there’s probably a difference between a state AG who fights exonerations in court and a governor who removes DAs from office.

Right. I don’t know what the political answer is. I do see a lot of hesitation in other states where the AG has a history of intervening — and there’s an even more pronounced hesitation to be vocal about getting an exoneration — to be out talking to the media about it. Letting the public know about the problems in the system is an important function of these units. But there’s a real reluctance to do that if it risks having the AG come in, take over, and move the case in the opposite direction.

But aside from the media component, these units are still doing their jobs. There was an exoneration in Florida just three weeks ago. So notwithstanding the fact that DeSantis is still governor, so there’s still a threat that he’ll retaliate, they’re still doing their jobs. They're still getting these cases done.

The last thing I wanted to ask you about concerns innocence. We have the Robert Roberson case in Texas, Marcellus Williams in Missouri*, Toforest Johnson in Alabama, the list goes on. These are all innocence cases in which Republican AGs are fighting an exoneration, in some cases over the objections of local elected DAs You’d think that not imprisoning or executing innocent people would be something we can all agree on. How did we get to the point where innocence has become part of the culture war?

[Note: This interview was conducted before Missouri executed Williams on September 24.]

I think part of the problem is that for most of these cases, we’ll never know the absolute certain ground truth. Even in DNA cases, there’s still often the slightest bit of gray area. So we still have this arrogance still that we’re getting it right all the time. I think do think we do get it right most of the time. But we get it wrong in a much higher percentage of the time than many are willing to admit. So you still get this arrogance. “No, no, no. This may happen in other places, but it doesn’t happen in my jurisdiction.” When I was doing this work Pennsylvania, I would say it was like the legislature felt there was some kind of magic pixie dust in the Allegheny Mountains that protected the state from what was happening in other states. They just didn’t think wrongful convictions were possible in their home counties.

So the problem is that even convictions for which the evidence overwhelmingly says we got it wrong, there’s almost always going to be just enough of a sliver of doubt for those people to say, “Nope, we got the right guy.” Maybe an eyewitness recantation came too long after their initial testimony. Maybe there’s some far-fetched explanation for the DNA results. There’s still going to a tiny of flicker of residual doubt there, and people who don’t want to believe the system makes mistakes are going to drive a Mack truck through it.

And then of course, you have the victims, who quite frankly are often treated horrendously throughout the entire process, including through the exoneration process. In too many jurisdictions they aren’t brought along as the CIU starts finding reasons to doubt a conviction. So this all comes out at once and the victims or their families are taken off guard. So they’re like, “Nope, this guy did it. We know that he did it.” And who's going to challenge them?

This is why John Hollway’s work on sentinel event reviews is so important. It’s so hard to get people to admit that this system they’ve devoted their lives and careers to can get something so profoundly wrong. Sentinel event reviews are about getting that institutional buy-in that there are flaws, that there are bad outcomes and that when they happen we need to figure out what went wrong and what we need to do to fix it.

The standard for a jury to convict is beyond a reasonable doubt. But once a conviction is set, you almost need that impossible “ground truth” you mention to overturn it. That process was established back when we thought the system always got it right. DNA testing is the closest we can get ground truth, and when that became available it showed that the system can make mistakes. But we never really course corrected in light of that revelation. If anything, it’s harder to get a conviction overturned now than it was then.

Right. The standard is supposed to be a preponderance of the evidence, but we know that’s not how it works. We know that the same judge who presided over a conviction is going to be much less likely to overturn that conviction than a judge who knows nothing about the case. But who do we sent these cases back to? The same judge who oversaw the conviction. The whole system is set up wrong.

I think it goes back to what I mentioned before — this unrealistic expectation that the system we have gets us to the truth. It doesn’t. It wasn’t even designed for that! I was a public defender for 10 years. I can tell you I did not go into court every day to find the truth. In some cases, my job was to obfuscate the truth. My job was to force the state to prove its own case, even when the state’s case was the truth. The whole reason we have cross-examination is to poke holes in the competing narratives. It’s not about truth.

Our system is designed to ensure fairness. Whether it does that well — and I think for the most part it does — that’s an entirely different thing. You can have a fair and constitutional trial that never gets to the truth.

But while we don't have a truth seeking system, but people think that we do. So CIUs are trying to adjudicate the truth within the confines of a system wholly unequipped for that. And that, I think, is at the heart of the disconnect.