What happens in our courtrooms and on streets changed on January 1, 2018. Alaska is using a new pretrial bail system unlike any other.

The state now relies heavily on-- and in some cases completely-- a computer-generated algorithm to make a decision about who stays in jail and who gets released back into the community before trial.

A little more than three months in, KTVA is taking an in-depth look at pretrial in Alaska: Where it came from, how it works, and why some say it doesn’t.


In 2014, Alaska's prison population was growing at a rate the state couldn't afford. The State Legislature created Alaska's Criminal Justice Commission through the passing of Senate Bill 64 and tasked its 13 members with finding ways to cut the prison population.

By 2016, Governor Bill Walker signed Senate Bill 91 into law.

The controversial omnibus bill was touted as a reform that would save the state hundreds of millions of dollars. SB91 increased the penalties for the worst offenses: first- and second-degree murder, and reduced sentences for a majority of the other crime categories, with the exception of homicide and sex offenses.

Public outrage, as crime continued to rise, leading to the passage of Senate Bill 54 in 2017, undoing some of what lawmakers decided went too far in SB91, all before Alaskans had seen its full effect.

Phase III of SB91, which deals with pretrial, began with the new year.

The commission relied heavily on The Pew Charitable Trusts when making recommendations, and pretrial reform was no exception. A brief from The Pew Charitable Trusts says research found “The number of pretrial inmates rose by 81 percent between 2005 and 2014, significantly outpacing the growth of the post-conviction population.”

People considered to be part of the pretrial population are individuals who have not been convicted of a crime and are legally presumed innocent.

Now, every person charged with a crime is evaluated under the state's new risk assessment tool. The tool produces a report that gives judges a computer-generated score between zero and 10 for a defendant, which indicates they're either low-, medium- or high-risk for two factors: how likely the person is to fail to appear for future court hearings, and how likely a defendant is to commit new crimes while out on bail before trial.

For each risk factor, the algorithm uses six data points to come up with a total score, then the higher of the two scores is what's given to judges. The data points include age at first arrest, the total number of prior FTA (Failure to Appear) warrants, and whether the person is currently booked on property or motor-vehicle (non-DUI) charges.

The tool’s data points are unique to Alaska, as more than 4000 statistical calculations of Alaska's criminal population data went into creating the algorithm.

A defendant’s score doesn't take into account factors like the nature of the charges in a defendant’s case, the possible sentence, potential harm to the public, weight of the evidence, a defendant's ties to the community or their personal financial resources.

What the tool cannot determine for judges is how dangerous someone might be-- yet, in cases in which a defendant receives a low or moderate score, depending on the charges in the case, the defendant could qualify for ‘Mandatory OR,’ which stands for mandatory own recognizance. If a defendant qualifies for mandatory OR, judges are prohibited by statute from setting a monetary bail.

“Mandatory OR meaning the judge would be obligated to release that person without setting a monetary bail. That's what mandatory OR means, they're released on the promise that they will come to court and they will follow the court's rules,” Robert Henderson, deputy attorney general in the Criminal Division, explained to members of the Senate Judiciary Committee in March.

Unfortunately, it doesn't always work out that way.

Jessica Malcolm, a felon, was arrested with a gun in January following a shootout in front of a trampoline park in the middle of the day on a Saturday. She scored a zero on the tool, was released the next day, and then fled the state, according to Bryan Schroder, US Attorney for the District of Alaska. 

In March, Shane Muse, a suspected two-time car thief who went on the run when he was supposed to show up in court for a forgery trial, scored a two on the risk tool, and also qualified for mandatory release, prohibiting the judge from setting a monetary bail.

“I understand that there's a criteria (sic) that's used by Pretrial Enforcement, but I'm here to tell the court that Mr. Muse presents a risk, a serious risk, to the public. That being said, if the court's hands are tied, then the court's hands are tied and you have to set bail the way that you're instructed by our new laws,” said Anchorage Assistant District Attorney Kevin Bergt during Muse’s arraignment.

Muse remained in custody, but only because another judge was able to set bail in one of his open 2017 cases-- which don’t fall under the new law.

Muse has several robbery convictions in the state of Hawaii, and Malcolm has a felony history in California, but that doesn't count here because the tool’s score doesn't consider out-of-state criminal history.

“We learned during the creation of the tool that out-of-state history was not going to be taken into account,” Henderson told lawmakers.

It turns out, the FBI would not share its data to set up the project. When asked by lawmakers why the Department of Law and the Department of Corrections moved forward with the tool anyway, Henderson said they had to because of SB91.

“The law was going to go into effect January 1, 2018, so we had to move forward with the tool,” Henderson said.

Henderson says one in three cases that qualify for mandatory release involve out-of-state criminal history. But even when out-of-state history isn’t present, KTVA has found cases in which police say a defendant on mandatory release has created more victims.

David Ramos doesn't have an out-of-state criminal history but was out on mandatory release for several January charges, including being a felon with a gun and vehicle theft, when Anchorage Police wrote in a criminal complaint that he was found in a stolen truck with someone else's bike, another man’s gun, a woman's purse, and another woman's groceries-- items representing five new victims.

“Several other states use a pretrial risk assessment tool, and several other states do not include out-of-state criminal history, but no other state has a mandatory release like we have,” Henderson said.

“Why did we do that?” the reporter asked.

“That is a policy question for the legislature,” Henderson answered, adding, “I will tell you, though, that is consistent with the recommendation that came out of the Criminal Justice Commission.”

Senator John Coghill (R-North Pole) is chair of the Senate Judiciary Committee and serves on the Alaska Criminal Justice Commission. He was instrumental in crafting SB91 and co-sponsored the legislation.

“We figured that we could mandate those to be released, but they could be released with accountability,” explained Coghill.

The accountability Coghill referred to is provided by the brand new Pretrial Enforcement Division (PED).

“What Alaska said is, ‘When you go out on bail, we're gonna monitor you,’ which we've never done before,” said Geri Fox, director of the PED.

Fox says she’s hired roughly 40 officers so far to operate the division she spent a year starting up.

PED officers are armed and have remand capability, meaning they have the authority to re-arrest a defendant who violates their conditions of release or commits a new crime.

A typical day for a PED officer usually starts with running risk assessments on defendants who have been arrested and will appear in court for arraignments in the afternoon, then fieldwork, which includes random compliance checks on defendants ordered to PED supervision.

“Our job is one-part law enforcement, one-part parent, and one-part social worker. Just like any parent, you want them to succeed. Just like any social worker, you want them to rehabilitate and never come back. And as a law enforcement officer, you don't want to arrest the same person over and over again,” PED Officer Chris Johnson said.

But this supervision isn’t guaranteed to every case. Of the 3382 people assessed in the first three months of the year, 1424-- less than half-- were assigned to PED supervision.

In fact, the pretrial risk assessment report for Ramos, who was deemed "low risk" with a score of five in the "new criminal arrest" category, specifies that PED supervision “is not recommended for defendants in this category.”

During the first two months of the year, defendants who were assigned to PED supervision had a one in four failure rate. Roughly 25 percent of defendants assigned to PED supervision between January 1 and March 1 either violated their conditions of release or committed new crimes, according to data from DOC.

KTVA also found that more than half of the defendants who failed on PED supervision were released again during a check of an inmate locator database in early March.

When approached with those numbers during an interview, Fox responded, “Defendants get out of jail, they have always gotten out of jail. Defendants get out of jail. In the State of Alaska, there is a constitutional provision that everyone has the right to bail.”

Fox later sent KTVA data indicating that roughly the same amount of people were being re-released before PED started and that the "re-release" rate even goes down, from 66 percent to 61 percent, for people under PED supervision.

“I think it's probably the single-best public safety element we had in our criminal justice reform program because it holds people accountable who had previously never been held accountable other than just a dollar amount, so I think pretrial has been a good thing,” said Coghill.

In the House, Representative Charisse Millett (R-Anchorage) isn't as hopeful.

“We’re putting Alaskans in danger every day when we continue to try and say that that pretrial risk assessment tool is working,” said Millett.

The Minority leader originally carried SB91’s companion bill in the House but ultimately withdrew her name and voted against it.

“When it came to looking for ways to put people in rehabilitation and getting people really the help that they needed, there was no front-end loading on SB91… It became a lot about saving money and not saving lives,” said Millett.

In the House Majority, Representative Andy Josephson (D-Anchorage) says he's also concerned with how Phase III of SB91 is playing out.

“Where I was open to reform was: I understand the inequality of saying, ‘You have $500 and have committed crime acts, I don't and have committed crime acts, and we have the same criminal history, and I can't get out of jail and you can.’ That, that needed fixing, but I think the pendulum got swung way, way too far,” said Josephson.

Facing vocal critics, dedicated supporters of the new pretrial system maintain that it’s working.

“If we don't want to assess risk, then okay. I guess it needs to be undone. If we don't want to have officers out there supervising these folks, then okay. I guess it needs to be undone. But, I will tell you this, we're making a safer Alaska by having these two pieces in place-- and to argue otherwise is just not being common sense (sic),” said Fox.

“It is, by all accounts, an experiment,” countered Millett, “And it is a bad thing to experiment on human lives.”

Opinions on the new pretrial system are widely different, but voices on both sides agree, Alaska’s fiscal crisis coupled with the opioid epidemic are other important factors to consider, along with the state’s laws, when looking at crime in Alaska.

An addict's perspective

Facing more felonies than he can count on one hand, "John Doe" is out on PED supervision, with an ankle monitor. KTVA agreed to protect this defendant’s identity in exchange for a candid conversation.

Reporter: “Are you addicted to substances?

Doe: “Yes.”

Reporter: “Which ones?”

Doe: “Um, I don’t discriminate… primarily heroin and meth.”

Reporter: “Would you say that’s fueled your cases as well?”

Doe: “I can truthfully say it’s the sole reason for my cases.”

Reporter: “Have you used substances since you were released?”

Doe: “I'm sure everybody on the ankle monitor uses substances. When you have an addiction, it's not just something you can give up overnight, you know?”

Reporter: “What do you want people to know about being an addict?”

Doe: “It's not always... every single person's got monkey on their back. It's just a... at different points in time in life, we choose to do different things. I made a mistake, you know? And I've been paying for it for a long time.”

He says the electronic monitor he was given doesn't help. He’s wearing the Transdermal Alcohol Detector (TAD). It tests his sweat for alcohol, which isn’t his problem.

"I can use heroin. I can use meth."

The device is not equipped with GPS capability. Instead, it operates using radio frequency, which measures how many feet it is away from the monitor, according to Doe. 

At the time of the interview, Doe said no one from the state had been to his home to check on the device’s power source.

“They might as well be giving you… it's a key to the city, for lack of a better way to put it,” said Doe.

Reporter: “Do you think the public thinks they're safer than they are?”

Doe: “Oh, most definitely. I know for a fact that they do. […] I felt a lot safer before I got on it.”

Reporter: “If you're arrested again for violating your conditions of release, what do you think the likelihood is that they'll let you right back out?“

Doe: “My understanding is the likelihood is 100 percent with the way that the new bail schedule works... It's a revolving door, it really is. They spit you out, to bring you back in, to spit you out, to bring you back in-- it's like, for what? Ridiculous. It's a joke.”

Electronic Monitoring

Following the interview with John Doe, KTVA took questions regarding the DOC’s use of electronic monitoring for pretrial defendants to both the DOC and a private electronic monitoring expert.

PED Director Geri Fox says the state is using four devices for pretrial electronic monitoring but has only confirmed that one of them has real-time GPS capability. That puts more than half of the state’s pretrial defendants on electronic monitoring on devices that cannot be tracked 24/7.

Fox declined to answer KTVA’s questions about Doe’s claims to be cheating his monitor, citing public safety.

“I run public safety, and I am not gonna tell defendants how those devices work,” said Fox.

KTVA made the decision to report Doe’s claims because it became apparent during research for this project that Doe’s is not an isolated case.

“I know that I’m being monitored here, and I know that those people aren't being monitored,” Troy Smith said during an interview at Alaska Pretrial Services, a private electronic monitoring company.

A couple years ago, Smith was, in his own words, “working” on his fifth felony. His crimes, also fueled by addiction, Smith found himself in a similar place as John.

He now credits Alaska Pretrial Services, the private company, for his success.

“It’s the longest I've been out since I was 15, 14-years-old,” Smith said.

He's under court-ordered 24/7 real-time GPS monitoring and an accountability plan the company made just for him, but Smith says he still has acquaintances who haven’t made as much progress and are now under the State’s new pretrial system, saying he’s seen them actively cheating their monitors “easily.”

Dennis Johnson, owner of Alaska Pretrial Services, says he is familiar with the state's devices but has never opted to put someone in the pretrial phase on a monitor without GPS.

“Active GPS is the only way to provide for community safety and victim safety and compliance,” said Johnson.

When it comes to GPS monitors, Johnson says not all models are the same, showing a steel cuffed GPS monitor a defendant shot at several times during a 17-minute attempt to get it off.

“We were able to get on scene and we actually took the individual into custody before he could get the device off his leg, and the device continued to report,” Johnson said.

He then showed KTVA what he says is the GPS device the state uses, and demonstrated his ability to cut through the strap with a pair of office scissors.

“That's unfortunate,” said Johnson.

In attempts to verify the strap Johnson cut was indeed similar to what the state is using, KTVA asked questions about the DOC’s GPS monitor, but Fox declined to answer them.

Fox did say that some defendants have cut their straps, but that the division is notified immediately when that happens, and also gets alerts if devices are tampered with.

She also noted the DOC is seeing success with their alcohol monitoring devices. 

“People are not drinking alcohol-- people who've been severely struggling with alcohol addiction-- so there's a story for every case out there, right?”

Johnson says he thought the DOC would contract with local companies who have been doing successful electronic monitoring here for years but says they were never approached specifically about pretrial.

The DOC maintains that they did, telling KTVA and lawmakers that they had several meetings with private companies, even supplying KTVA with a log of emails, calls and in-person meetings.

“The way the private EM (electronic-monitoring) individuals were making a living was having people pay for the service,” DOC commissioner Dean Williams explained. If you couldn't pay for the service, guess what? You stayed in jail. So for me, as commissioner, that disparity bothered me and the law made it clear that my job was to get people out, regardless of whether they're a poor person or not.”

Johnson says if the state contracted with them, it would remove financial responsibility from defendants, but the DOC went with the out-of-state company BI Incorporated.

“The Department of Corrections, with this Pretrial Enforcement Division, has operated in competition with the private sector. We had 12 companies when this all started. There's now four left in the state. And we're not far behind,” said Jonson.

Through a Freedom of Information Act Request, KTVA obtained a copy of the state’s contract with BI Incorporated. The contract does not distinguish services for pretrial versus sentenced individuals.

According to the contract, the device John Doe has cost the state roughly $6 per defendant per day, while the kind of GPS monitoring services Johnson wanted to provide the state would have run DOC about $15 per defendant per day.

Proposed Solutions

Both critics and supporters of the new pretrial system and risk tool agree that tweaks are needed, but how far those adjustments should go is where the debate continues. 

Mia Costello (R-Anchorage) championed SB91 in the Senate, but now she is leading the charge in the capital to repeal it.

The Republican senator from Anchorage surveyed Alaskans about crime and found the response overwhelming.

“A good friend’s car was stolen and the thieves drove around doing drugs,” said Costello, reading a survey response from a constituent.

She has two large binders full of stories like that one and believes the only way to regain Alaskans' trust is to un-do SB91 completely with her measure, Senate Bill 127.

“It was a massive undertaking and I was a co-sponsor of the legislation, as were many of my colleagues, and we got it wrong,” said Costello.

In the House Majority, Representative Andy Josephson (D-Anchorage) doesn't see a way forward for a full repeal.

“It’s a little bit like a tanker leaving port, you can't turn it on a dime. I wish it had not happened. I was the only member of my party to vote against SB91 in 2016. I fought it with every ounce of my being. But it did, and I don't know that we can just pretend it never happened,” said Josephson.

House Minority leader Charisse Millet (R-Anchorage) supports a full repeal of SB91, but she's also sponsoring House Bill 335, a measure to make changes to the risk assessment tool.

“It says that mandatory release is off the table. And it's a tool, and it's not the rule. Even if someone scored low on that, the judge still has the discretion.”

Millett says she doesn’t want there to ever be a situation where a judge thinks someone is dangerous but cannot keep them away from the public.

“I believe that is happening now,” she said.

The least invasive proposal for changing pretrial comes from Governor Bill Walker's desk and seems to be getting the most traction. The initial version of SB150 would allow judges to override the tool when it calls for mandatory release, but only in cases in which the defendant has an out-of-state criminal history.

That bill got its first hearing in the Senate Judiciary Committee on March 21 after a March 9 email to the judiciary chairs from Attorney General Jahna Lindemuth. The email cites a KTVA story in which two alleged repeat car thieves qualified for mandatory release and urges them to hold hearings on the Governor's bill, saying, "It may be a permanent fix to this awful situation.”

Copyright 2018 KTVA. All rights reserved.