The Layperson’s Guide To Why Stenographic Reporting Is More Efficient Than Digital Reporting & ASR

Though I have written very much on this topic, never have I sat down and written something that can be handed to anybody. Today’s the day! This is the context of the stenographer shortage forecasted eight years ago and ongoing debate regarding that. It is my hope that having this out there lets court reporters share this instead of having to painstakingly debate the merits one discussion at a time. If you’re looking for a career as a stenographic court reporter, check out National Court Reporters A to Z program, Project Steno, and Open Steno. My resource page also has a lot of industry links to help find manufacturers, retailers, and educators.

Digital Reporting

In the court reporting and captioning industries, there are three modalities, stenographic, voice writing, and digital. Stenographic reporting has a stenographer using a chorded keyboard called a stenotype to write strings of letters that form words, sounds, and statements. Voice writing consists of speaking into an automatic speech recognition system trained to the voice writer. Digital reporting consists of having a person record the testimony, ideally with multitrack or sophisticated equipment, monitor the audio, and annotate it for a transcriber. Some companies, like Verbit, insist that by running audio through an automatic speech recognition system, technology can “improve the human.” There is a fraud being perpetrated on consumers and digital reporters themselves just about every day where they are being told these models are equivalent or even that digital reporting is better. Let’s go deeper.

It’s not cheaper. Where the stenographic reporter works on a per-page basis with a small appearance fee, the digital reporter is paid hourly, typically within the range of $15 to $50 an hour. The cost of the transcript is added to that, which is heavily subsidized off the exploitation of offshore transcribers, who can make as little as $0.84 a page or $0.24 a minute, as I learned in my previously published discussion with one. The companies utilizing digital reporting are trying to cut their transcription costs down to a fourth of what they were, but none of these savings, if any, are being passed to the consumer or courthouse. There have also been past studies that show samey costs for digital and steno. When one factors in the cost of replacing equipment, which likely has a useful life of around four years, the cost for digital can only spiral out of control. We stenographers eat all of our equipment costs. In the event that we were deleted from the field, the lack of competition could only drive those recording equipment costs up further (higher demand for ostensibly the same supply). In addition, reporting companies would have no incentive not to pass those costs to the consumer because there would be zero competition to keep them in check.

Storage is more expensive. This often gets waved away because of the “common sense” mantra that digital storage is CHEAP. But again, this is a cost that is often taken up by the court reporter and would likely transition to the court and consumers. Audio files can be anywhere from 10 to almost 10,000 times larger than our raw stenographic text files. This means that in a worst-case scenario, courts are spending 10,000 times on storage what they would if it was a stenographically reported court. In a best-case scenario, a digitally recorded court could switch to stenographic reporting and bring their storage costs down to about one ten thousandth of what it is today. I arrived at these numbers by comparing hundreds of my raw stenographic notes for 7-hour days to the average recording rates of various formats. Those files are usually in the ballpark of 200 to 400 KB or 28 KB an hour. Something like a an mp3 can be 28 MB an hour. Something like a wav can be 10 MB a minute! To put these numbers into perspective, an MB or megabyte is about 1,000 times larger than a KB or kilobyte. Technology fans point to YouTube and its ability to compress and hold so many videos, to which I have always pointed out that YouTube is a for-profit business, so unless we want to sit through advertisements while waiting for our court cases, maybe they’re not a good model for justice system record storage. We are also assuming that Youtube would share its compression with the government at no additional cost. I have also pointed out that YouTube has existed for less than 17 years. Some courts have record retention requirements that exceed 50. So assuming that the digital model would translate well in perpetuity is shortsighted and a leap of faith. It’s much more likely to provide spiraling costs and force shorter record retention times on courts in the future.

It’s less efficient. By switching to digital reporting, a company is necessarily adding to the number of people required to do a job. Math-wise, theoretically, we were always bound to be 2x to 4x as fast because average typing speed is in the ballpark of 50 words per minute, really fast transcribers get to 100 words per minute, and stenographers are taking down stuff at 225 words per minute. My theory was completely wrong; it requires many more people or much more time. Self-reported, transcribers can take up to six hours to transcribe one hour of audio. According to a Lisa Migliore, as revealed in my PCRA article, the company Verbit even admitted that it would take up to eight transcribers to provide a next-day transcript. Stenography is constantly belittled for its low pass rate. The argument is made that we cannot solve the shortage because of an 80 to 90% dropout rate. But if we’re going to multiply the number of people required by 8 or 9 (someone should man the recorder according to AAERT best practices) then this argument is hollow and an obvious lie. Let’s say you have 10 stenographic reporters. You get 100 students, and 90 percent drop out, you’ve replaced those 10 stenographers. If you need 9 people to replace one stenographer, then we’re already talking about 90 people to replace those 10 stenographers with digital. If you can recruit and train 90 people, why are they not being recruited and trained for stenography? Anybody that says we cannot solve the shortage is saying “we can recruit 90 people, but recruiting 100 is impossible.”

Court Reporting Industry Outlook 2013-2014, Ducker Worldwide

The turnover is comparatively astronomical. Stenographic court reporters stay in the business past retirement age. You can see this from a snapshot of our field’s age about eight years ago, pictured above. The average employee stays about 4 years. The stenographic reporter stays three or four decades. So in addition to requiring about the same number of recruits to fill demand, we know that getting them to stay means getting them into stenographic reporting. Failing to use stenographic reporters means needing those same 90 people I mentioned above every 4 years. Again, assume 10 stenographers retire. We can recruit 100 students who will replace those stenographers, and in 12 years there is a high likelihood they’ll still be in the game. Alternatively, we can recruit and train about 100 digital reporters / transcribers who will stay about four years, and then another 100, and then another 100. So if you’re willing to recruit 300 people over 12 years but not willing to recruit 100 people over the next 2, you’re not thinking logically. 40 years, those same 100 recruits produce those same 10 stenographers. Meanwhile, statistically, one will have needed to recruit and train about 900 digital court reporters and transcribers.

No education culture exists for digital reporting. As pointed out in a past article, when looking at comparable nonprofits, stenographic nonprofits and our education system has about ten times the funding and five times the number of schools. This lack of support on the digital side will allow for companies to dictate best practices to digital reporters and transcribers, and corporate morality is highly questionable. There’s already evidence that cost shifting is being used to kill competition. What happens when we move from a field where there is already guidance on what to do when asked to alter the record to one where an employee is seen as interchangeable and low skill? How much pressure will those poor people be under to change things if they’re making a fourth of what we are? This happens. I’ve been asked to change stuff. Colleagues have been asked to change stuff. Digital reporters and transcribers will be asked to change stuff and they will be under incredible pressure to actually do it. We can stop that by letting them know about steno. As of eight years ago, stenographic court reporting was said to be some 96% of the field. It makes perfect sense for digital reporters to join the majority and become stenographers.

Taken from the Ducker Report, page 6.

Companies are already showing contempt for their own digital reporters and that will likely only get worse. As I pointed out in a past article, while advertising digital reporters to lawyers, Verbit put out an infographic that stated they were not highly trained. There’s already this misconception that digital reporters are not skilled workers, which will make it easier for them to be exploited, to the aforementioned detriment of the consumer. Even with all of our stenographic institutional knowledge and support systems, companies managed to freeze the value of our rates for over three decades. If companies are successful in pushing the narrative that digital reporting is the future, there’s no reason to believe they won’t freeze the digital rates right where they are and the cycle of exploitation will continue. Think I’m wrong? Stenographic court reporters are already being offered below half of market rate for their services. So there’s a SHORTAGE and companies are doing everything in their power to make stenography look as unappealing as possible. Great way to make the situation worse. If we cannot count on them to not exploit people during a shortage, we cannot count on them. This also heavily incentivizes digital reporters to become stenographers, because there are nonprofits to help set them up with mentors in almost every state. Not so with digital reporting.

Offshoring will put transcribers beyond the reach of our courts. There has been no secret that digital reporting companies are offshoring their work to places like Manila, Kenya, or India. In the stenographic model, if there is a problem with the transcript, the stenographer can be subpoenaed to testify about the stenographic notes he or she took down. In the digital reporting model, the digital reporter may be available to testify as to what he or she can recall, but the transcriber will likely be outside the reach of our law enforcement. Add in the fact that there might be eight transcribers to ask about any one transcript, and we have a major problem. Just from experience, I can tell readers that when you’re listening to hundreds or thousands of proceedings, it all starts to blend in. Without someone contemporaneously taking down stenographic notes at the time of a proceeding, how can one be sure audio has not been tampered with? Audio is easy to edit without any sophisticated tools. Advanced technology like deepfakes and voice cloning will make tampering that much easier. There are already enough fakes in the world that forensic audio analysts are necessary. Taking court reporters out of their role of guardian of the record will only add to the chaos and lead to an era of “to the best of my recollection, that sounds like what happened.” The legal field needs to avoid declaring this modality adequate. People’s lives and livelihoods are on the line and it’s painfully obvious that there’s been no real thought put into shifting from stenography to digital reporting. What good is blockchain-like tracking if you can’t get a hold of everyone involved in the process?

Centralization encourages attacks and successful attacks would be more catastrophic. As I have seen, there are two distinct ways that courts operate. Either they operate on a centralized model where the court controls all the records or a decentralized model where reporters are given limited autonomy to store and edit their records. Some courts run on a hybrid model where they keep the raw stenographic notes in case of a problem but the reporter is responsible for the maintenance of all else. Something like grand juries in New York City would be a centralized model. The district attorneys buy the stenotypes and equipment and all work is done on the office’s intranet, which has limited connectivity with the outside world. Decentralization is more along the lines of having each reporter responsible for maintaining their records. Stenographic court reporting leans toward decentralization, meaning that anyone looking to sabotage court records in bulk would have to hack not only the court’s system but also each individual stenographer. In a centralized model, a system compromised by a steganography attack or some other attack could be wiped or hijacked. Basically malicious pieces of code can be hidden inside images and audio, and court computers can then be used to mine Bitcoin for criminals. Decentralization typically suffers from lighter security but makes it harder for a would-be attacker to get access to everything, especially machines that are not connected to the internet. Centralization typically has amazing security, but once an attacker is in, they have access. In a hybrid court like many of New York’s courts, a gang of criminals would have to break both the decentralized and centralized models to wipe records. Good luck with that! It’s no wonder that the New York courts have told ASSCR President Eric Allen many times “we want stenographic court reporters!” Again, we take on the cost of our equipment and the responsibility of its security independent of the system. This is a working condition we proudly accept because it creates redundancy and reduces the risk of lost records. This problem by itself wouldn’t be a big deal, there’s some risk no matter what, but combined with all else is pretty damning.

Lost audio eats any savings. It is always a bad day when any record goes missing. But again, taking this model from one where the stenographer is responsible for the record and can land themselves in extremely hot water if they do not do what they are supposed to do and trusting an audio-only record is dangerous. As I posited in the Summer 2020 NYSCRA Transcript, just three hours of reconstruction hearing can cost $1,000 in resources. In the best-case scenario where people are following best practices and an audio monitor is present and responsible for the audio, this risk is less. But as I’m about to show you, there’s no comfort there either.

Consumer protections for stenographers are better. I would happily assume those audio monitors take their job as seriously as I do. But considering that digital reporting companies have shown no qualms with offshoring or even posting legal proceeding audio on the internet, it’s time to put this to bed. Proponents of digital court reporting are only going to follow best practices where they are forced to by law. Putting our faith in that is foolish. It’s been over a hundred years since the stenotype was invented and regulations related to our field are still incredibly relaxed in some states, like New York. I’d wager that at this rate it’ll take another hundred years before legislators get around to protecting digital reporting consumers and courts. Consumer choice is the way to win here. If people demand a stenographic reporter every time, we can start focusing on how to improve ourselves rather than having to justify our existence every time somebody with a smartphone goes “I can do your job!”

Even in states with incredible consumer protection, companies that support digital reporting, like US Legal, have argued they’re not bound by those laws because they are not court reporting companies. Yet 70% of their business is court reporting? How much more obvious does it need to get that these companies don’t care what is right or legal? They care only what they can get away with. And they got away with violating the law in the Holly Moose case seemingly because plaintiff had trouble showing damages. The government’s position is clear: “You can violate our laws, claim you’re not providing the service that makes up the bulk of your business, and be as dishonest as you want to be. As long as the wrong person is calling you out on it, it’s just fine by us.” Just to give this some more context, the California licensing board for reporters has a long history of being as anti-court reporter as possible. They refuse to step in and regulate digital court reporters, but their stenographic reporters have some of the most stringent regulations in the country. The licensing board of California is basically tying two arms and a leg behind every stenographic court reporter’s back and stenographic reporters are still winning the race there. This is probably a result of bureaucracy and not some conspiratorial nonsense, but it’s the reality that stenographic reporters are living — governments are quick to come after us and can’t be bothered with enforcing the law against anyone that can afford a legal team. Assuming arguendo that digital proponents’ position that the modality is equivalent to stenography is true for just a moment, it’s almost like here in New York City if we decided it was important to license barbers that use scissors and strictly test them on their skills and knowledge, but having no regulation if they practice as a barber using an automatic shaver/buzzer. Both could mess up your hair pretty bad in the wrong hands. Not too many people would want only one to be under government-granted license. Now throw that same idea down, but instead of hair we are talking about appeal records that could decide whether or not someone goes to jail, gets executed, or loses a million dollars.

Just to be clear for anyone that believes the incompetence of regulatory authorities is overstated or hyperbole, according to the California Regulatory Law Reporter, Spring-Summer 1990, authorities have known about digital reporting, then called electronic reporting, for thirty years. Now they’ve expressly disclaimed their obligation to protect consumers. It’s writ-of-mandamus levels of incompetence. A licensing board so useless that court reporters and consumers would be better off asking the California legislature to dissolve it in its entirety.

Provided to me by Protect Your Record Project/Kelly Bryce Shainline

Likely financial suffering of companies using digital reporting. In a 2019 Kentley Insights report, it was noted that 1 in 4 court reporting companies is not profitable. If the vast majority of the field is stenographic and only companies that can afford to run at a loss are the ones attracting big private equity money, it follows that the unprofitable minority are the ones that are trying to change the field are doing so at a loss. They are able to do this because if they are successful the rates can subsequently be jacked up and those losses can be recouped. This is something that was labeled as anticompetitive decades ago, but go ahead and try to prove such a thing in court — maybe the court will once again decide the plaintiff is missing some obscure element and therefore any and all misbehavior is just fine! This also has a regular-world example. Uber did its best to kill the taxi industry at a loss to itself. Then its prices skyrocketed up 50% or more. Want your court reporting bill to double in the future? Allow these companies to continue to push digital reporting.

Blatant dishonesty of companies that push digital reporting. In addition to my thoughts about worker exploitation and the Holly Moose issue, I’ve tackled more misconceptions in the last five years than anyone should ever care to. Just some examples, VTestify had a calculator on their site that claimed they could save thousands per deposition, which was false. Veritext was trying to train attorneys to allow for digital court reporting in deposition notices before ostensibly throwing its then VP of Sales under the bus for going in the exact direction the company was going. US Legal stuck a bogus equation on JD Supra to make the stenographer shortage seem hopeless. Verbit says whatever makes it look good at the time. This is who we want to entrust the future of legal records to? First there was a question of whether digital reporting was even in use by some of them. That was then thoroughly documented and there was a shift to “we can’t hire stenographic reporters due to shortage” which was a lie because few if any recruitment efforts were made to find stenographers through public directories like NCRA Sourcebook. What kind of excuse is that? How about I walk into court tomorrow and say “I’m really trying, judge, I just can’t get the words.” If these companies cannot figure it out they are useless to the legal profession and should dissolve so that stenographers can meet the needs of their clients. It’s become difficult to wave away the sheer incompetence and failure to assist stenographers as honest mistake. If I give you a flat tire once and apologize, chances are you’ll forgive me. If I do it several times over half a decade, how long are you turning the other cheek? At what point do you turn around and tell me “stop now or I will take action to make you stop”? We have suffered too many flat tires. We are ready to go as far as this needs to go. The law itself recognizes that, in dealing with human affairs, there are very few things in this world that we know with absolute certainty. Anyone waiting for an admission of dishonesty better stop holding their breath. In many ways, Stenonymous has become the force to make this stop. We were pushed and pushed to a point where the path of least resistance was to punch back with what we’re best at, preserving history. I’m working on a couple of articles about US Legal’s dishonesty in particular. When I’m done with them, I’ll link them here.

Automatic Speech Recognition.

Inherent danger to already-marginalized speakers. When the Testifying While Black study came out, news media slapped court reporters around a bit because we were only 80% accurate when being tested on African American Vernacular English dialect. I slapped back. The study showed we were twice as good as the average person at understanding this dialect and 1.5x as good as lawyers. As the New York State Court Reporters Association told the New York State Unified Court System, use of automatic speech recognition technologies would not threaten access to justice, it would implode it. To do away with stenographic court reporters would be doing away with the people that, with no formal dialect training whatsoever, understand AAVE speakers the most. We are not a monolith, and some of us will always perform better than others, but as I’m about to show you, we outperform computers too.

Doesn’t actually work. Companies like Verbit have backpedaled from their stance that automation is the way to go to saying things like “technology will enhance the human.” They’re convinced having human transcribers clean up ASR output is the way to go. With all the major players having difficulty getting accuracy above 80% in the study “racial disparities in automatic speech recognition,” it’s clear why. There are inherent problems surrounding ASR through machine learning and the constant presence of language drift. It seems highly unlikely that stenographic reporting or captioning could be adequately automated. Have you seen a really good ASR demo? Please consider that just like Peter Molyneux and Project Natal, that demo might be scripted. I demonstrated how easily voice recognition could be faked by faking it myself with my amateur computer coding knowledge. If all the richest companies in the world are having extreme problems with automatic speech recognition and making it consistently perform at high accuracy, it follows that the small fish like Verbit, Otter, or Parrot are not being honest about their capabilities. Some people marvel at the technology today. “I can just dictate this into my phone! We might not need you guys soon, heh, heh, heh.” Yes, the technology has come along wonderfully. Would you like me to have Siri or Alexa transcribe your loved one’s murder trial? No? Oh…

If it cannot be automated, then we are talking about having all the problems inherent to digital reporting.

Some Related Myths

Over the past few days I’ve had it pointed out to me that some of the corporations I’m accusing of being anti-steno send financial support to Project Steno or help with schools. As for the schooling, with half of our industry being in New York, Illinois, Texas, and California, if a school is not training students for those places, it is a drop in the bucket and meaningless compared to the measures being taken to crush us. As for supporting nonprofits, there’s probably a write off there, so it’s not quite as altruistic as one might assume. And then that leads to, “well, it’s a business.” Exactly. Now you get it. It’s a business. It does not care about you, me, my field, this country, or anything else. The only thing that makes it care is when we unequivocally tell it it will cease to exist if it doesn’t start caring now.

How About This?

Debate and reason are vital to the health of our field. Confronting uncomfortable truths has been nothing but victory after victory. Let’s keep on the road of information dispersal and advocacy. Most people will support us if they are armed with the facts and circumstances. We’re at a tipping point, and the way that the cards fall depends largely on how hard we blow. Learned something new? Pass this along and help increase consumer critical thinking so that they catch this stuff before it gets to this point. See a flaw in my reason? Point it out. Worst that happens is I abandon a weak position for a stronger one.

Addendum:

I should say that this is all omitting the possibility that something is inaudible on a recording or that a recording goes missing. Careful consideration will have to be given to who is responsible and to what extent when things don’t work right if the country is to adopt a full digital model. Stuff happens even to the best of us. The benefit of the stenographer is that he or she can alert parties as soon as there’s an equipment issue. In a world of full automation, which is where digital proponents clearly want to go, parties would get to find out about problems whenever they ordered the transcript, which could be days, months, or years after a proceeding.

5 thoughts on “The Layperson’s Guide To Why Stenographic Reporting Is More Efficient Than Digital Reporting & ASR

  1. I already thought you were a legit journalist before this, but ok… another great piece, Bro….keep ’em comin’….

Leave a Reply