With so many trial and appellate courts now using Zoom to live-stream their hearings during this historic pandemic, I wondered if this partnership could be a vehicle for courts to real-time close caption their live-streamed hearings and arguments.
Spoiler alert answer: Not yet.
I downloaded the Supreme Court of the United States’ oral argument audio file in its recently decided case Ramos v. Louisiana. I uploaded that audio file to Otter.ai to have a written transcript prepared. (I was not expecting perfection on that front.) You can see and hear the output below.
I also downloaded the typed transcript from the Supreme Court of the United States. (Again, I was not expecting perfection.)18-5924_Ramos-transcript-1
My question was: Would the Otter.ai typed version of the audio at least be good enough to link up with Zoom so that Otter.ai could provide real-time notes (perhaps closed captioning) of a judicial hearing or argument?
I didn’t preformulate any metrics to make a yes-or-no determination. While the Otter.ai’s typed version is particularly good in many ways (it nailed the Hurtado and Balzac references), there are other areas that convince me, at this moment, it’s still not yet the right tool for real-time court hearing closed captioning. Everyone: keep your court reporters and pay them well.
Again, Otter.ai got many things correct. But other things gave me pause.
Three items that took the most Otter.ai hits were:
- Justice Powell
- caselaw reference to “Apodaca” and
- stare decisis
And there were many other things that gave me pause. The first item shows what Otter.ai recorded. The [bracketed item] reflects the phrase in the SCOTUS transcript:
justice Powells [Justice Powell's] justice Powles just as Powells justice Powell his justice panel's of justice opinion justice pals just spell justice pebbles justice powers justice scorches [Justice Gorsuch] Just as Briar [Justice Breyer] justice semi or [Justice Sotomayor] just a story [Justice Story] it doesn't happen [Justice Kavanaugh] Justice Cavanaugh [Justice Kavanaugh] appa Dhaka [Apodaca] APA DACA APA data Napa DACA after DACA APA Doc APA Dhaka epidemic data APA doc a APA doc Have APA docket avid not the doctor starry decisis [stare decisis] story decisis Star decisis start decisis sorry decisis vicinity [vicinage] know it [note at] know [note] note about [what about] status has [state has] alternate [I'll turn] pastures prudence [past jurisprudence] Tim's [Timbs] data stano opinion [DeStafano opinion] then Shiro again Summerlin [in Schiro against Summerlin] 12% rule [12-person rule] 12% requirement [12-person requirement] the developers requirement [the 12-person requirement] unity [unanimity] they tend to verdict [a 10-2 verdict] commission [community] get attend to verdict [get a 10-2 verdict] get a six years [get a 6-0 verdict] jury recessed decision [jury reached its decision] board coke [Lord Coke] for one for reasoning [4-1-4- reasoning] no stare decisis the fact [no stare decisis effect] if the penitentiary [if a petit jury] we thought to be difficult [would start to be difficult] how reasonable there are alliances [how reasonable the reliance is] like in pain you read where he goes [like in Pena-Rodriguez] that I entered and clear with the court [that I am clear with the court] note [amendment] deliberate use [deliberations] that staff [that statute] and may just be clear [and let me just be clear] reason settled until a dolphin [It was unsettled until Apodaca] even Emily [unanimity] the faithful being powerful [the fifth vote being Powell's vote] desk camps [Descamps] was mural. [Ms. Murrill] they will for justice, as you said, [there were four justices who said] that seems to me the mission that we are [that seems to me an admission that we are] It precedent waste or anything [If precedent weighs for anything] And why does an estate take that [And why doesn't a state take that] to a court [to accord] non unanimous jury vote Nixon [non-unanimous jury verdicts and] It's a strategy [Its structure] said in solving [said in Sullivan] which is red face [which is read] believe our on our song [believe are on our side] role [rule] word enemas [were non-unanimous] we fought 25 briefs [we filed 25 briefs] the Federal right in the country. institution that does include [the federal right in the constitution does include] T [Teague] Careful underlined centrists [Counsel, on your reliance interests] anyone's little Liberty entrusts [anyone's liberty interests] if you accept the legal privacy's? [if you accept the legal premises] weren't those [why aren't] six granted union and union Animal robots [six, granted, unanimous rule but] And secret tend to verdict [and seek a 10-2 verdict] rule from the lawyer against Hogan [rule from Malloy against Hogan]
A final observation. If you take a few moments to watch the video of the Otter.ai transcript played against the audio, Otter.ai was not always intuitive in recognizing different speaking voices. [And sidenote: the light green highlights are mine.]
I use Otter.ai for lower-stakes matters and I know that it’s a valued tool for many reporters. Otter.ai has some helpful utility. But it’s not quite there yet for live courtroom closed captioning.