*The initial version of this post was written in a vulgar, ranting, raving, and confusing manner due to a medical episode I was having. Sorry for how that was written. I have scrapped it. It held no creative or intellectual value that cannot be reproduced at a later date. Many of the people I named and cursed at are actually good people that I admire.
See the new version below:
This time, with love.
The “conspiracy” is simple. It is very similar to what Purdue Pharma did to healthcare. Lies and misdirection, as well as possible government corruption. I should note that in our case, because the government has not yet investigated to my knowledge, it is hard to tell if we’re seeing tacit parallelism or a fully-planned plot. That said, numbers do not lie, and there are problems with the numbers.
Veritext, US Legal, STTI, and others publish misleading and demonstrably false info. There is likely a baseline assumption that it won’t matter in 10 or 20 years because AI will “take over.”
Our associations, either by design or habit, just keep staying the course. Old tricks don’t work in a new and modern world. I start to change minds with a $10,000 budget and become suspicious of the association management industry. Dave Wenhold, whose intelligence and charisma I openly admire, becomes suspicious to me. This is about where I start to deteriorate. How do I use my media to bring people to associations that shut down new ideas without debate; as NCRA did for Amendment 5 this year? How do I support associations that give responses so boneheaded a six year old could do better? Even my beloved NYSCRA seems intent on crushing new ideas quietly and behind closed doors.
NYSCRA also seems oddly inert whenever I suggest anything that would strengthen the association.
Then, like with Purdue, there is potential government corruption. There is a concept called the revolving door, where someone on behalf of the company catches the ear of a government official and the government official just so happens to do what the company wants. Then the government official retires to a nice new job with the company. So far I’m told California Court Reporter Licensing Board officials have actually done this — which is incredibly suspicious. California may well be the state with the strictest regulations for court reporting. It refuses to regulate digital court reporting. Consumers should probably demand the complete dissolution of the board. It would be like heavily regulating snack foods but refusing to regulate pretzels.
Let’s not forget Arizona, where my public comment was completely ignored by Director Byers, even though I sent him an e-mail well before my personal issues began. Very convenient for AAERT and anyone who thinks courtrooms should be investing in recording equipment.
There are also issues here in New York. It is apparent that members of the public are being told no spots are open and the government is sitting on provisional applications.
I cannot say for sure what we are seeing. I do see a pattern of passive aggressiveness toward court reporters from corporations, associations, and government — a passive aggressiveness that peculiarly fades any time someone questions it.
The antitrust conspiracy may not be an actual conspiracy. But if I was designing a plot to exaggerate and exacerbate the stenographer shortage, it would look exactly like what we are experiencing today.
A source that shall remain anonymous passed me information that NYPTI prosecutors were being given education about stenography and court reporting that doesn’t match up with reality. During the presentation, stenography was made to look as old and outdated as possible when in actuality it is top-of-the-line tech in speech-to-text transcription. This supports my belief that the stenographer shortage is being intentionally exaggerated and exacerbated.
Stenographers are no strangers to social media. We’ve had students like Isabelle Lumsden get thousands of eyes on our stenotypes. We have amazing content from accounts like Stenoholics. More recently, I got to see a video from the TikTok letsgetfries. The video starts with our hero mentioning that she’s been on jury duty for two weeks. The most important thing she’s learned? Stenographers have the wildest energy of anyone she has ever met in her life! Don’t fuck with them. Maybe she’d make a good court reporter, she got our hand and eye thing down already!
This is exactly how I look at work every day for the last 11 years and she nails it.
I bring this up for the entertainment value, but also as a reminder that strategically social media is our battleground. There are companies out there right now, like US Legal, that are claiming the stenographer shortage cannot be solved by training more stenographers. It’s a blatant lie dressed up like industry news to fool industry insiders and outsiders. Meanwhile, we know from the Open Steno 2021 Survey that about two thirds of people coming into contact with steno, at least in that community, are coming into contact with it thanks to the Internet. So we’ve got to out-presence them, recruit people, and steer our students clear of dishonest companies.
OH NO. EQUATIONS. RUN.
And make no mistake that I am calling US Legal dishonest. In their article they note 1,120 retirees a year and 200 new reporters. An annual shrinkage of 920 reporters, giving the impression that this is an annual gap that never ends and only gets larger. But that’s not how these numbers work. First of all, they’re extrapolated from the Ducker Report, which was a forecast based off 120 interviews and some proprietary data analysis, not a future-telling machine. As more and more reporters retire out, the retirees would decrease each year. Anybody with a second-grade math level can figure out their math is wrong because a shrinkage of 920 annually means there would be zero reporters in 30ish years. That’s not actually possible if you’re getting 200 new reporters a year. The equivalent would be me going on JD Supra and saying the CEO of US Legal gets two brain cells a year and loses ten, therefore his company will probably be bankrupt in ten years. Doesn’t matter if it’s true, it just sounds good. I don’t begrudge people for where they work, but as a company, no matter how great any individual employee might be, they’ve got to be among the most dishonest, toxic, harmful companies in our industry. You know that scene in Star Wars where Luke tells Kylo “amazing, every word of what you just said was wrong”? That’s how I feel. Reporters get some cognitive dissonance here because US Legal does have nice people working for them, but that doesn’t change how I feel about the entity itself. It’s like Theranos. I’m sure nice people worked there, but the entire operation was a big joke that should never have happened.
Letsgetfries, I don’t know if you’ll ever happen across this, but let’s just say we’re so used to being treated like potted plants that whenever anybody says anything nice about us, we boost them big time. From getting Stanley Sakai’s article featured on Medium last year or sharing John Belcher’s deposition strategies. You’re no different. As of late last week we had shared you over a thousand times! Hope you had a great experience with jury duty! If you know anybody who’d like to join our field, please let them know about NCRA A to Z, Project Steno, or Open Steno. For the record, our crazy energy is mostly thanks to everyone saying they can replace us and failing for the last half decade. We’re working it out. Thanks again!
Spreading through social media is a clip from John Belcher. He talks about how he got his dream job as a prosecutor, which allowed him to be in court almost every day and work with court reporters and other court staff. He talks about all the things that court reporters hope attorneys talk about. Some key takeaways?
Don’t do something you wouldn’t do in front of the judge. They read the transcripts.
Don’t step on the witness. Count to four before starting the next question or answer.
Speak a little slower. He suggests 70% speed.
Don’t disrespect opposing counsel, the witness, the court reporter, or other attendees.
Be careful about side discussions that take away or distract from the proceeding.
Adding fillers at the beginning of questions like “okay” or “perfect” may create bad habits for trial questioning.
Preparation is key. Expecting the court reporter to put up your exhibits for you may burn valuable time.
Don’t take it from me, check out his video on LinkedIn today! You can also see his YouTube here.
A close friend sent me a Bill Maher clip from a while back. Obviously, Maher has his political leanings, but after he gets done with flaunting those, he makes a decent point. He describes the over-engineering of society and gives some pretty striking examples. His preferred vape’s newest model has no mouthpiece despite being something you put in your mouth. Car handles are replaced with buttons in some cars despite no efficiency gains. He describes a situation where his rental car asked him if he’d like to open the trunk while going 60 miles an hour. The point is clear, change for the sake of change is not always worthwhile or efficient. Indeed, change for the sake of change can be very dangerous.
This is connected to the exaggerated claims of salespeople that I’ve written about extensively, especially as it relates to voice recognition. I described it several posts ago as the claim game. Anybody can say anything. Anybody can make their business seem like the new, hot thing. Take this blog post by Kaplan Leaman & Wolfe from about a year ago. It reads nicely, and it sounds innovative. It mentions a flat-rate fee, affordable per-page price structure, a design to significantly reduce legal expenses. At the point in 2020 the post was written, everybody was doing remote stuff. Pretty much everybody’s got a per-page price structure. Anybody can claim their service is affordable or reduces expenses. It’s called puffery and it’s an ordinary part of business.
Where it gets messy, and where I’ve tried to educate reporters, is some advertisements are easier to spot than others. If Burger King says they’ve got the best burger, most everyone knows that’s puffery and sales. Things get harder with technology. How do you prove or disprove whether someone has made a technological breakthrough without a comprehensive understanding of the science and concepts at work? Not all reporters understand the concept of machine learning. Even those of us that have researched quite a lot can’t possibly know everything there is to know. This leaves a gap for tech sellers to come in and try to fool consumers into buying services that may not suit their needs using the hype train.
Told you I write a lot about this. I read a decent amount too.
This also leaves reporters playing a catch-up game of learning about these systems so they can help their clients navigate claims and discern fact from fiction. For example, the truism that technology is improving every day. We look around ourselves and marvel at this magical modern world. But I’ve taken the pretty hard stance that certain technologies, namely voice recognition and associated technologies, are not improving every day. Give it speech it’s used to and it’ll do fine. Give it speech that’s just a little off from what it’s trained for and it’ll turn “would you raise your right hand” into “it’s rage right hand.”
Yes, it’s rage’s right hand.
But surely reinventing the wheel and all these claims of being BETTER aren’t BAD for business, right? If puffery is normal then a little bit of stretching the truth won’t hurt anybody! But we already see that’s not the case. Take Maher’s example. One little glitch on the highway and you could have dead motorists. Take the fact that 25 percent of court reporting companies may be unprofitable; court reporting has been around a long time, it’s likely the losers are the ones trying to switch it up too much too fast. Take vTestify’s massive switch from boasting about providing inexpensive court reporting services to providing an online platform for the legal industry. Take Verbit’s claims in its series A funding of 99 percent accuracy and its subsequent announcement that it will use human transcribers after all, and the very real possibility that it is, despite all its funding, not profitable.
Exaggerated claims serve only as a cliff from which these companies have a chance to walk off of or step back from. The competition is going to wise up. The consumers are going to wise up. I can only hope that a lot of these tech companies realize this, wise up, and start putting their resources behind actually improving our technology. It’s a lot easier to compete in a field with maybe seven players like Stenograph or Advantage than it is to beat out thousands upon thousands of independent contractors and hundreds of reporting firms, many with their own clients and connections. It’s frighteningly easy to see there’s a more lucrative path than over-engineering what stenographic court reporters have made simple, and I can only hope that business owners realize this before walking investors’ money off that cliff.
This month I had a chance to sit down with Marc Russo of MGR Reporting. Marc’s a working reporter and business owner. We got to hit a lot of topics in this video, including Marc’s history in the field, how reporter skill relates to reporter treatment, and how scheduling ahead can help reporting firms fill their clients’ needs.
Using Marc’s words, it’s about treating reporters like people instead of numbers.
Very often on stenographer social media, we get questions about whether something should be reflected as said, sic’d, or “corrected.” There has been plenty of discussion over the years on whether to correct lawyers’ or witnesses’ speaking in transcription. There are a lot of ways to take this conversation, and in the spirit of keeping this fun, I’ll hit the highlights.
Necessary in this discussion is: “What is my transcript?” The bulk of freelance work goes to deposition reporting. When a case is filed and initial motions to dismiss are decided, if the case is not dismissed, it moves to discovery. Discovery is where the parties exchange information that they have so that when it is time for trial, there are few or no “surprise” pieces of evidence. At the conclusion of discovery, the parties can ask the court to decide the case as a matter of law if there are no factual questions in dispute. If the case cannot be resolved as a matter of law, it goes on to trial. An integral part of the discovery phase is deposition testimony. Parties have an opportunity to question the other side’s witnesses under oath. Witness testimony is evidence, and the evidence unveiled during the discovery phase is ultimately what helps parties settle cases, courts decide whether a matter can be decided as a matter of law, impeach witnesses at trial, and appellate courts review the decisions of the trial court. In America, the testimony of one witness can convict beyond a reasonable doubt. Your transcript is the verbatim record of what occurred during the testimony, and again, that testimony is powerful evidence.
Unsurprisingly, there are many different takes on what “verbatim” means. We can all read the dictionary definition: “in exactly the same words that were used originally.” But court reporting and transcription are service industries, and there have been many times where court reporters are pressured by a client or company to change that verbatim record in some small way. In my view, that pressure gave life to a lot of court reporter conventions that are daunting for students, new reporters, and even veteran reporters to master. For example, as a young reporter, I was told to take out false starts, never ever report “um,” and to even physically remove strikes and withdrawns from deposition transcripts. Now, wherever you are, the laws in your jurisdiction supersede my advice or opinion, but I am going to share the way I look at each in the hopes that this can be shared with others who struggle with these. For sure, anything I write can and will be debated, but debate can only improve our field.
Removing False Starts
This was drilled into me by agencies as a young reporter. “Always remove false starts.” It’s still being pushed on young reporters today, to the point where some may not even be taking them down. Frankly, I see this as bad advice. The essential factors for a reporter to consider in the way something is transcribed are context and readability. Does my transcription of the verbatim notes change the context of this testimony? Does my transcription degrade the readability of this testimony? In my view, removing most false starts will not actually change context, and they will improve readability. As an example:
“Q. Are you — did you go to the store?”
“A. Yes.”
It would be difficult to argue that removing the words “are you” and simply changing the question to “Did you go to the store?” hurts the context. Nothing has changed. And so to the extent removing false starts is looked at favorably in our field, I get it. But what about when it would change context?
“Q. Are you — I mean, did you go — did you go to the — sorry. Did you, if you remember, go to the store?”
“A. I’m sorry. I don’t understand your question.”
What happens in a world where a young reporter, told that they must remove false starts, removes all that and changes it to “Did you, if you remember, go to the store?” The context is unequivocally changed. Verbatim, it’s very clear that the question was not clear. There was a lot of extra “stuff” in there. If such a question is cleaned up, it makes the witness look like they’re not paying attention or unintelligent. Removing false starts can hurt the context and stop legal professionals from doing their job. Imagine that the deposition is taken by a young associate and the trial lawyer is a seasoned vet who did not sit on the deposition. Reading a “cleaned up” version, the trial lawyer might believe the witness is a bumbling mess. When that witness gets on the stand and is given clear questions, it’s going to be a surprise for that trial lawyer. So even where law may allow the removal of false starts, it’s a decision the court reporting practitioner should make using their own sound judgment, and not on the whims of an agency or client. You may also want to see NCRA Advisory Opinion 4 to the extent it touches on this topic.
Never Ever Report Um
Again, I see the reporting of “um” as a matter of context and readability. Let’s say that you’re taking a motion argument, and it looks something like:
“MS. ATTORNEY: Um, um, um, um, um, um, um, um, um — your Honor, based on the hearing that we just had, there is no set of facts under which the people may prevail. I therefore ask you to dismiss this case in the interest of justice.”
Does it really change anything if you don’t report the ums in that specific instance? Nope. And this isn’t a hypothetical. I recall a situation just like this, where the attorney had, without question, made the point they were trying to make, and then became very flustered asking the court to make a decision. But what if the situation was a trial situation?
“Q. Did you see Mr. Vanhorten shoot Mr. Gorfasi?”
“A. Um, well — um, yes.”
If you transcribe that sentence as “well, yes” the context is destroyed. The witness seems crystal clear on what they saw. Those ums have a kiloton of context that transform what is being said. I’m not here to say anyone who omits an um is a bad reporter, but think twice before subscribing blindly to the “truism” that we do not report ums.
Physically Remove Strike That or Withdrawn
Often, strike that is seen as a false start. Just imagine the typical scenario:
“Q. Were you — strike that. Were you ever an employee of ABC Corporation?”
Again, the rule of context comes into play. In the above scenario, I can’t say I see a big problem with the omission of the false start strike that. But as a mentor to many over the years, I’ve come across the following scenario:
“Q. Were you ever an employee of ABC Corporation?”
“A. Well, I wasn’t an employee at the time.”
“MR. GUY: Move to strike.”
What have mentees come back and said? “Chris, my agency says remove strikes. Do I remove that whole thing?” Working reporters have had to counsel many a new reporter. “No. We cannot remove portions. That motion to strike is the attorney preserving their motion on the record, which will be later reviewed by a court.”
Ultimately, with these three categories, leaving things in as they are said is often the way to go. A court can always seal, strike, or disregard something that shouldn’t be in the transcript. On the other hand, a reporter that does not put something in the transcript can be questioned about why it was removed, or even have their neutrality called into question.
Mispronunciations
Now that we’ve explored some of the common things that impact context, let’s explore some more “what ifs.” Since I was a newbie, the discussion has come up, “Someone said a word incorrectly. Should I sic this?” This comes from a very literal way of thinking sometimes cleverly but pejoratively termed in our field as “the literati.” The pressure is turned up to make something “perfectly verbatim” when there is a video, which brings up the question “are we not being verbatim when the video camera’s not on?” There are two major schools of thought, literal verbatim and readability, and within those schools of thought, you have many different situations and many different gradients. I could not possibly address each one, but let’s hit some common examples.
“Let me ax you a question.” It’s obvious to anyone that the speaker means to say ask. Many speakers do not enunciate clearly. It does not change the context to transcribe “ask,” and it greatly improves the readability, so for such moments where the context is not endangered and the word is obvious, there’s no harm in having the correct word rather than some kind of phonetic spelling. I would say the same for names. Let’s say someone’s name is Dr. Giglio. One person says “Jig-lee-oh” and the other says “Gig-lee-oh.” Again, if it’s clear that this is the same person, and the context is not endangered, transcribing the correct name is the way to go. If it’s not clear, then it’s time to speak up and get some clarification on the spelling! This is not to say you can never write a name phonetically, but try to make these spellings consistent throughout the transcript to the extent people are saying the same word, even if they say it a little differently.
“It’s supposably true.” In addition to not changing context by being too verbatim, we have to be mindful that sometimes people use words that sound like other words. If someone says a “wrong” word or a word we are not accustomed to hearing, we must resist the urge to correct, because that actually can alter context. We must also take the time to research things we are not a hundred percent sure on. In my book, supposably was not a word. The WordPress spellchecker says it’s not a word. I came to learn, a decade into my career, that supposably means “as may be conceived or imagined.” Supposedly is more of a synonym for allegedly. Was this true 10 years ago? I have no idea. As court reporters, we face the harsh reality of language drift. Words fall in and out of use. People do not speak as we were taught. So while you might correct something like axing a question, you have to think twice before you correct something that’s “supposably wrong.” If you have three minutes, check out my favorite video illustrating language drift. You can go back about 700 years before English starts sounding like gibberish and giraffes were camelopards. Through a mix of self-initiated research and our continuing education culture, we keep ourselves ahead of the average transcriber.
Whether there is video or not, you want a clear and logical reason why you have transcribed something the way you transcribed it. In my view, the strongest reason for a transcription choice is “transcribing it any other way would change the context or was not verbatim.” Reporter convention and training take a backseat to that.
What devilry is this?
Dialects
Court reporters are masters of English dialects even when we have no training. There is a study out there that pretty much shows we are twice as accurate as laypeople when transcribing the AAVE dialect. The thing that makes us, as humans, so much better than computers at transcribing speech that has a dialect or an accent is our ability to understand context. For example, in the Northern Cities Vowel Shift dialect, someone might say something that sounds like “she went down the black.” Dependent upon the context, we know that that sentence can be “she went down the block.” In brief, our ability to look at the totality of a statement is important. What a reporter may hear is “down the black.” But what must be transcribed, in the interest of both context and readability, is “down the block,” unless there’s some context that tells us “black” is actually correct.
This is also where our ability to speak up for the record comes into play, because if a reporter is unsure, they can seek clarification. For purposes of our work, dialects and accents are very much like garden-path sentences where a sentence goes in a different direction from what you were anticipating; we can discern what’s said from the context. Though accents are a different animal from dialects, the same rules apply. Early in my career, I had a gentleman say something that sounded like “I got up and leave her.” Through context I knew the statement was “I gotta pull a lever.” He was explaining how to open bus doors! Another man talked about the “zeh bruh lies or stripes” on the road, which could only be “zebra lines or stripes.” We’re not here to pick apart how something was said, we’re here to take down what was said.
Latin
“Vice-a versa” versus “vice versa.” “Neezy preezy” versus “nisi prius.” “Nun pro tunc” versus “nunc pro tunc.” “In forma papyrus” versus “in forma pauperis.” Because of Latin’s considerable history and various modern regional pronunciation schemes, this is another thing that gets confusing fast. My advice? Treat it like mispronunciations. Treat it like dialects. Treat it like all these other examples and look at the context. If someone says, objectively, the wrong phrase, then don’t change it for them, but if you know exactly what they said, don’t transcribe it phonetically for the sake of “verbatim.” Take a look.
“MR. GUY: Quid pro quo is the Latin phrase for ‘from possibility to actuality.'”
So we head over to Google, and we can see clearly that “a posse ad esse” is the Latin phrase for that. Quid pro quo means “something for something.” No correction is necessary here. We knew what was meant, but the wrong thing was said. Verbatim is our friend. But what if it’s just a butchered pronunciation?
“MR. GUY: vee-low-shee-yee-yus quam asparagi coke-a-tor is the Latin phrase for ‘faster that asparagus can be cooked.'”
MR. GUY: velocius quam asparagi coquantur is the Latin phrase for ‘faster than asparagus can be cooked.'”
If you’re following along, you can probably tell that I think the second one is the obvious choice. No matter how butchered that pronunciation might be, if it’s clear, transcribing the wrong word or a series of phonetic jabs is what a computer would do. You’re better than that, use it to your advantage. And do not be too hard on yourself for making a mistake. I have had colleagues that were told the incorrect spelling of Latin phrases by people far more educated than many of us are. Whatever the issue, learn from various mistakes and situations, try not to become so rigid with regard to language that it endangers context, and continue to grow.
But I Was Taught This Way
Whenever stuff like this comes up, inevitably you’ll get responses like “but I was taught this way,” or “I’ve been doing it my way for 30 years.” Nobody can really fight with that. We have to respect one another and those various perspectives, backgrounds, and experiences. But I’ve come to look at it from a liability and reputation perspective for the freelance court reporter. If someone questioned you on a transcript, how would you respond? “My agency told me to” is a very unsafe response, because the agency can just say they didn’t, and if you’re an independent contractor, they’re not supposed to have direction and control over you. So take a look at the practice, and imagine being questioned on it. “That’s what you said” is a much stronger response than “everybody does it this way.”
We have to deal with the fact that, while we may live in a world of “truisms,” like “clients expect us to clean up the record,” these things are not universal, and in fact, as a young reporter, I had a lawyer tell me “you can’t change [false starts], it’s part of the record!” Imagine being about 20, and repeatedly told that “everyone cleans it up,” “this is normal,” “this is expected,” “you’re a bad reporter if you don’t fix it,” and then being slammed with “you can’t take that out.” It’s not surprising to me that there are reporters of all ages and experience levels that struggle with this. I’m really hoping this helps the strugglers: I was you. You’re not going to have an immediate answer for every situation, but having an objective or neutral method for how you make these decisions is imperative. If problems arise, and they occasionally do, you’re going to be defending your work. Remember, this is all about having an accurate record for review by the parties, trial courts, and appellate courts. Our expertise is what stops errors like “lawyer dog” from making it into the record and ruining people’s lives. If your work hasn’t changed the context of a statement and the transcript is readable, you’re off to a great start.
Allie Hall is a reporter and educator who has made amazing strides in getting schools to pick up court reporting programs and getting students filling those programs. Some months ago, a group of working reporters came together under Allie’s guidance and leadership, and with additional help from co-admin Traci Mertens, the group has managed to donate thousands to new reporters and students in need.
If you are a working reporter or CART writer looking to give back, please reach out about joining the group. There is a fundraiser currently ongoing, and working reporters may donate ten to twenty dollars to help meet students’ needs.
Working reporters may donate via:
Venmo: Allison-Hall-89
PayPal: allie441@gmail.com
Google Pay: allie441@gmail.com
There is truly no contribution too small. If you’ve got an extra ten dollars to put down on a student, consider sending it along to Allie today! I am a contributing member of the group, and I have rarely ever seen such energy and accountability in a grassroots fundraiser. This is something special, it’s something I really support, and I know the money is going to making the road that young professionals have to travel just a little bit less bumpy. Most of us can look back at our student years and say “I wish I had…” Now we get to be a part of making sure the students of tomorrow have!
There’s a lot of conjecture when it comes to automatic speech recognition (ASR) and its ability to replace the stenographic reporter or captioner. You may also see ASR referred to as NLP or natural language processing. An important piece of the puzzle is understanding the basics behind artificial intelligence and how complex problems are solved. This can be confusing for reporters because in any of the literature on the topic, there are words and concepts that we simply have a weak grasp on. I’m going to tackle some of that today. In brief, computer programmers are problem solvers. They utilize datasets and algorithms to solve problems.
What is an algorithm?
An algorithm is a set of instructions that tell a computer what to do. You can also think of it as computer code for this discussion. To keep things simple, computers must have things broken down logically for them. Think of it like a recipe. For example, let’s look at a very simple algorithm written in the Python 3 language:
Do not despair. I’m about to make this so easy for you.
Line one tells the computer to put the words “The stenographer is _.” on the screen. Line two creates something called a Stenographer, and the Stenographer is equal to whatever you type in. If you input the word awesome with a lowercase or uppercase “a” the computer will tell you that you are right. If you input anything else, it will tell you the correct answer was awesome. Again, think of an algorithm like a recipe. The computer is told what to do with the information or ingredients it is given.
What is a dataset?
A dataset is a collection of information. In the context of machine learning, it is a collection that is put into the computer. An algorithm then tells the computer what to do with that information. Datasets will look very different dependent on the problem that a computer programmer is trying to solve. As an example, for enhancing facial recognition, datasets may be comprised of pictures. A dataset may be a wide range of photos labeled “face” or “not face.” The algorithm might tell the computer to compare millions of pictures. After doing that, the computer has a much better idea of what faces “look like.”
What is machine learning?
As demonstrated above, algorithms can be very simple steps that a computer goes through. Algorithms can also be incredibly complex math equations that help a computer analyze datasets and decide what to do with similar data in the future. One issue that comes up with any complex problem is that no dataset is perfect. For example, with regard to facial recognition, there have been situations with almost 100 percent accuracy with lighter male faces and only 80 percent accuracy with darker female faces. There are two major ways this can happen. One, the algorithm may not accurately instruct the computer on how to handle the differences between a “lighter male” face and a “darker female” face. Two, the dataset may not equally represent all faces. If the dataset has more “lighter male” faces in this example, then the computer will get more practice identifying those faces, and will not be as good at identifying other faces, even if the algorithm is perfect.
Artificial intelligence / AI / voice recognition, for purposes of this discussion, are all synonymous with each other and with machine learning. The computer is not making decisions for itself, like you see in the movies, it is being fed lots of data and using that to make future decisions.
Why Voice Recognition Isn’t Perfect and May Never Be
Computers “hear” sound by taking the air pressure from a noise into a microphone and converting that to electronic signals or instructions so that it can be played back through a speaker. A dataset for audio recognition might look something like a clip of someone speaking paired with the words that are spoken. There are many factors that complicate this. Datasets might be focused on speakers that speak in a grammatically correct fashion. Datasets might focus on a specific demographic. Datasets might focus on a specific topic. Datasets might focus on audio that does not have background noises. Creating a dataset that accurately reflects every type of speaker in every environment, and an algorithm that tells the computer what to do with it, is very hard. “Training” the computer on imperfect datasets can result in a word error rate of up to 75 percent.
This technology is not new. There is a patent from 2000 that seems to be a design for audio and stenographic transcription to be fed to a “data center.” That patent was assigned to Nuance Communications, the owner of Dragon, in 2009. From the documents, as I interpret them, it was thought that 20 to 30 hours of training could result in 92 percent accuracy. One thing is clear: as far back as 2000, 92 percent accuracy was in the realm of possibility. As recently as April 2020, the data studied from Apple, IBM, Google, Amazon, and Microsoft was 65 to 80 percent accuracy. Assuming, from Microsoft’s intention to purchase Nuance for $20 billion, that Nuance is the best voice recognition on the market today, there’s still zero reason to believe that Nuance’s technology is comparable to court reporter accuracy. Nuance Communications was founded in 1992. Verbit was founded in 2016. If the new kid on the block seriously believes it has a chance of competing, and it seems to, that’s a pretty good indicator that Nuance’s lead is tenuous, if it exists at all. There’s a list of problems for automation of speech recognition, and even though computer programmers are brilliant people, there’s no guarantee any of them will be “perfectly solved.” Dragon trains to a person’s voice to get its high level of accuracy. It simply would not make economic sense to have hours of training a software to everyone who is going to speak in court forever until the end of time, and the process would be susceptible to sabotage or mistake if it was unmonitored and/or self-guided (AKA cheap).
This is all why legal reporting needs the human element. We are able to understand context and make decisions even when we have no prior experience with a situation. Think of all the times you’ve heard a qualified stenographer, videographer, or voice writer say “in 30 years, I’ve neverseen that.” For us, it’s just something that happens, and we handle whatever the situation is. For a computer that has never been trained with the right dataset, it’s catastrophic. It’s easy, now, to see why even AI proponents like Tom Livne have said that they will not remove the human element.
Why Learning About Machine Learning Is Important For Court Reporters
Machine learning, or applications fueled by machine learning, are very likely to become part of our stenographic software. If you don’t believe me, just read this snippet about Advantage Software’s Eclipse AI Boost.
Don’t get out the pitchforks. Just consider what I have to blog.
If you’ve been following along, you’ve probably figured out, and it pretty much lays it out here, that datasets are needed to train “AI.” There are a few somewhat technical questions that stenographic reporters will probably want answered at some point:
Is this technology really sending your audio up to the Cloud and Google?
Is Google’s transcription reliable?
How securely is the information being sent?
Is the reporter’s transcription also being sent up to the Cloud and Google?
The reasons for answering?
The sensitive nature of some of our work may make it unsuitable for being uploaded. To the extent stuff may be confidential, privileged, or ex parte, court reporters and their clients may simply not want the audio to go anywhere.
Again, as shown in “Racial disparities in automated speech recognition” by Allison Koenecke, et al., Google’s ASR word error rate can be as high as 30 percent. Having to fix 30 percent of a job is a frightening possibility that could be more a hindrance than a help. I’m a pretty average reporter, and if I don’t do any defining on a job, I only have to fix 2 to 10 percent of any given job.
If we assume that everyone is fine with the audio being sent to the cloud, we must still question the security of the information. I assume that the best encryption possible would be in use, so this would be a minor issue.
The reporter’s transcription carries not only all the same confidential information discussed in point 1, but also would provide helpful data to make the AI better. Reporters will have to decide whether they want to help improve this technology for free. If the reporter’s transcription is not sent up with the audio, then the audio would only ostensibly be useful if human transcribers went through the audio, similar to what Facebook was caught doing two years ago. Do we want outside transcribers having access to this data?
Our technological competence changes how well we serve our clients. Nobody reading this needs to become a computer genius, but being generally aware of how these things work and some of the material out there can only benefit reporters. In one of my first posts about AI, I alluded to the fact that just because a problem is solvable does not mean it will be solved. I didn’t have any of the data I have today to assure me that my guess was correct. But I saw how tech news was demoralizing my fellow stenographers, and I called it as I saw it even though I risked looking like an idiot.
It’s my hope that reporters can similarly let go of fear and start to pick apart the truth about what’s being sold to them. Talk to each other about this stuff, pros and cons. My personal view, at this point, is that a lot of these salespeople saw a field with a large percentage of women sitting on a nice chunk of the “$30 billion” transcription industry, and assumed we’d all be too risk averse to speak out on it. Obviously, I’m not a woman, but it makes a lot of sense. Pick on the people that won’t fight back. Pick on the people that will freeze their rates for 20 or 30 years. Keep telling a lie and it will become the truth because people expect it to become the truth. Look how many reporters believe audio recording is cheaper even when that’s not necessarily true.
Here’s my assumption: a little bit of hope and we’ve won. Decades ago, a scientist named Richter did an experiment where rats were placed in the water. It took them a few minutes to drown. Another group of rats were taken out of the water just before they drowned. The next time they were submerged, they swam for hours to survive. We’re not rats, we’re reporters, but I’ve watched this work for humans too. Years ago, doctors estimated a family member would live about six more months. We all rallied around her and said “maybe they’re wrong.” She went another three years. We have a totally different situation here. We know they’re wrong. Every reporter has a choice: sit on the sideline and let other people decide what happens or become advocates for the consumers we’ve been protecting for the last 140 years, before the stenotype design we use today was even invented. People have been telling stenographers that their technology is outdated since before I was born, and it’s only gotten more advanced since that time. Next time somebody makes such a claim, it’s not unreasonable for you to question it, learn what you can, and let your clients know what kind of deal they’re getting with the “new tech.”
Addendum 4/27/21:
Some readers checked in with the Eclipse AI Boost, and as it was relayed to me, the agreement is that Google will not save the audio and will not be taking the stenographic transcriptions. Assuming that this is true, my current understanding of the tech is that stenographers would not be helping improve the technology by utilizing this technology unless there’s some clever wordplay going on, “we’re not saving the audio, we’re just analyzing it.” At this point, I have no reason to suspect that kind of a game. In my view, our software manufacturers tend to be honest because there’s simply no truth worth getting caught in a lie over. The worst I have seen are companies using buzzwords to try to appease everyone, and I have not seen that from Advantage.
Admittedly, I did not reach out to Advantage myself because this was meant to assist reporters with understanding the concepts as opposed to a news story. But I’m very happy people took that to heart and started asking questions.
As a stenographic court reporter, I have been amazed by the strides in technology. Around 2016, I, like many of you, saw the first claims that speech recognition was as good as human ears. Automation seemed inevitable, and a few of my most beloved colleagues believed there was not a future for our amazing students. In 2019, the Testifying While Black study was published in the Language Journal, and while the study and its pilot studies showed that court reporters were twice as good at understanding the AAVE dialect as your average person, even though we have no training whatsoever in that dialect, the news media focused on the fact that we certify at 95 percent and yet only had 80 percent accuracy in the study. Some of the people involved with that study, namely Taylor Jones and Christopher Hall, introduced Culture Point, just one provider that could help make that 80 percent so much higher. In 2020, a study from Stanford showed that automatic speech recognition had a word error rate of 19 percent for “white” speakers, 35 percent for “black” speakers, and “worse” for speakers with a high dialect density. How much worse?
The .75 on the left means 75 percent. DDM is the dialect density. Even with fairly low dialect density, we’re looking at over 50 percent word error rate.
75 percent word error rate in a study done three or four years after the first claim that automatic speech recognition had 94 percent accuracy. But in all my research and all that has been written on this topic, I have not seen the following point addressed:
What Is An Error?
NCRA, many years ago, set out guidelines for what constituted an error. Word error guidelines take up about a page. Grammatical error guidelines take up about a page. What this means is that when you sit down for a steno test, you’re not being graded on your word error rate (WER), you’re being graded on your total errors. We have decades of failed certification tests where a period or comma meant a reporter wasn’t ready for the working world yet. Even where speech recognition is amazing on that WER, I’ve almost never seen appreciable grammar, punctuation, Q&A, or anything that we do to make the transcript readable. It’s so bad that advocates for the deaf, like Meryl Evans, refer to automatic speech recognition as “autocraptions.”
Unless the bench, bar, and captioning consumers want word soup to be the standard, the difference in how we describe errors needs to be injected into the discussion. Unless we want to go from a world where one reporter, perhaps paired with a scopist, completes the transcript and is accountable for it, to a world where up to eight transcribers are needed to transcribe a daily, we need to continue to push this as a consumer protection issue. Even where regulations are lacking, this is a serious and systemic issue that could shred access to justice. We have to hit every medium possible and let people know the record — in fact, every record in this country — could be in danger. The data coming out is clear. Anyone selling recording and/or automatic transcription says 90-something percent accuracy. Any time it’s actually studied? Maybe 80 percent accuracy, maybe 25; maybe they hire a real expert transcriber, or maybe they outsource all their transcription to Kenya or Manila. Perception matters; court administrators are making industry-changing decisions based on the lies or ignorance of private sector vendors.
The point is recording equipment sellers are taking a field which has been refined by stenographic court reporters to be a fairly painless process where there are clear guidelines for what happens when something goes wrong, adding lots of extra parts to it, and calling it new. We’ve been comparing our 95 percent total accuracy to their “94 percent” word error rate. In 2016, perhaps there were questions that needed answering. This is April 2021, there’s no contest, and proponents of digital recording and automatic transcription have a moral obligation to look at the facts as they are today and not what they’d like them to be.
