Ex Machina: Technological Disruption and the Future of Artificial Intelligence in Legal Writing

• Autore: John Campbell
• Carica: Co-director of the Denver Empirical Justice Institute at the University of Denver Sturm College of Law (US), where he is also a member of the Lawyering Process faculty
• Pagine: 294-326

†John Campbell is the co-director of the Denver Empirical Justice Institute at the University of Denver Sturm

College of Law (US), where he is also a member of the LawyeringProcess faculty. He gives particular thanks to

Dr. Catherine Durso of the University of Denver for her work on the statistics presentedin this paper. Without

her,this article would not be possible. He also thinks Kensye Woodand Megan Largson, his researchassistants,

for their dedicated work.

llV

OLUME

5llI

SSUE

2 A

RTICLES

& E

SSAYS

U

NIVERSITY OF

B

OLOGNA

L

AW

R

EVIEW

hps://doi.org/10.6092/issn.2531‐6133/12238

A

Ex Machina: Technological Disrupon and the Future of

Arcial Intelligence in Persuasive Legal Wring

J

OHN

C

AMPBELL

†

ABSTRACT

Technology is disrupting the practice of law and revolutionizing how lawyers work. This

revolution is made more powerful because it is increasinglycoupled with a rigorous and scientific

approach to the law. In some ways, law is looking more like a Silicon Valley startup and less like

the oak-paneled law firms of the last 200 years. As law, technology, and science merge, the

implications for the profession are wide-sweeping. This article explores persuasive legal writing,

offering new thoughts on what the future will hold. Specifically, this article pilots a method for

applying technology and science to measure, analyze and improve persuasive legal writing,

offering it as a proof of concept that anchors the article’s broader, and perhaps more

controversial assertion. Namely, more powerful and refined persuasive legal writing software

tools, fueled by artificial intelligence, should and will disrupt and reshape significant portions of

the legal space, including how legal writing is taught and how it is produced. The effect will be to

view legal writing as more science, and less art. The next set of luminaries won’t rely on anecdote

or intuition to teach or create legal writing; they will rely on software and data.

KEYWORDS

Legal Technology; Legal Research;Legal Writing; Empirical

294

2020] UNIVERSITY OF BOLOGNA LAW REVIEW [Vol. 5:2

TABLE OF CONTENTS

Introduction ............................................ 295

1. Technological Revolution in Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297

1.1. Examples from Other Legal Settings . . . . . . . . . . . . . . . . . . . . . . . . . 297

1.2. Revolution in Legal Research and Writing Too . . . . . . . . . . . . . . . . . . . 299

2. Demystifying Persuasive Writing through Technology . . . . . . . . . . . . . . . . . . . . 303

2.1. What the Legal Writing Experts Say . . . . . . . . . . . . . . . . . . . . . . . . . 305

2.2. Early Empirical Research Regarding Legal Writing . . . . . . . . . . . . . . . . 307

2.3. Focus on Judicial Writing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 308

2.3.1.Intensifiers.................................... 309

2.3.2.Readability.................................... 310

3.ThePilotEmpiricalStudy.................................... 312

3.1. United States Supreme Court Briefs are Simpler . . . . . . . . . . . . . . . . . . 316

3.2. Writing Style Correlates to Winning . . . . . . . . . . . . . . . . . . . . . . . . . 319

4. Improving the Method and Tools Available . . . . . . . . . . . . . . . . . . . . . . . . . . 321

5. Implications for the Profession in the Future . . . . . . . . . . . . . . . . . . . . . . . . . 322

Conclusion ............................................. 325

INTRODUCTION

Empirical legal writing studies, powered by new technologies, will fundamentally disrupt

and revolutionize how we think about persuasive legal writing. This article argues,

through a pilot study that serves as a proof of concept, that, over the coming years, the

fog of advice about persuasive legal writing style can largely be cleared by developing

better tools to measure persuasive legal writing and better methods for studying the

effect of legal writing on outcomes. It argues that, as persuasive legal writing becomes

more science and less art, legal writing software powered by artificial intelligence1will

disrupt a variety of fields, including how legal writers create briefs, the legal insurance

industry, legal finance of cases in litigation, and howlegal writing is taught to students.

We can and should move away from anecdote and assumption, and towards

software and data. For years, professors (including me), legal writing gurus (like Bryan

Garner who has made millions teaching legal writing), and judges (like the late Antonin

Scalia) have talked about “how” to write effectively. But the truth is that this advice is

largely untested – and as some law students would gladly tell you – too often

inconsistent.

1A.I. is an umbrella term that coversa range of technologies that learn over time as they are exposed to more

data. P

EDRO

N

AVA ET AL

., A

RTIFICIAL

I

NTELLIGENCE

:AR

OADMAP FOR

C

ALIFORNIA

(Little Hoover Commission 2018),

https://lhc.ca.gov/sites/lhc.ca.gov/files/Reports/245/Report245.pdf. A.I. is the quality of any computer

system (data, algorithms, analytics, bots, etc.) the ability to sense, reason, adapt, learn, and understand just

like humans can. Id. In the deeply developed sectors, A.I. technology can encompass the ability to reason

through to conclusions and learn to adapt specific outputs or behaviors to circumstances. Id.

295

EX MACHINA

For years, I have told my law students that if you were to put ten world-renowned

appellate lawyers in a room, gave them the same legal problem,and then had them write

briefs, those briefs would be profoundly different. To be fair, some of this variation is a

result of options in the legal analysis sphere. But assume for a moment you gave the

attorneys the same cases, the same strategy,and the same frames. Assume youeven gave

them the general order in which the arguments would be presented. The briefs would

diverge nonetheless. The writing style among the briefs would vary, sometimes

significantly. To fully understand this notion, you only need to read a few briefs for the

United States [hereinafter U.S.] Supreme Court. World-renowned attorneys with proven

track records take markedly different approaches in their written advocacy, so much so

that their styles are a sort of fingerprint and the author can be guessed by the style and

tone alone.

The same variance occurs if you put ten legal research and writing professors in

a room. Within my own department, we have engaged in rigorous debates about how

students should be taught to write during our monthly meetings. We all think we know

the “truth” about what lessons or writing methods work. The picture is no different if you

read books on legal writing, or if you attend conferences dedicated to the art.

Why are there such marked disagreements? The pat answer is that legal writing

is complex, and personal preferences dominate. We might even say authors have to be

themselves, and find the approach that works for them. I take a different view. We see

variance in what people treat as “good” legal writing because we suffer from a severe

deprivation of data. Like ancient people performing rain dances because their

understanding was limited by the availability of knowledge, we argue about what

“works” in persuasive legal writing because we do not actually know. Sure, we have

hunches. And to be fair, some of these are formed over years of experience,making them

more like educated guesses or very crude statistical inferences. But there is little hard

data upon which to draw any salient conclusions. And absent hard data, the best we can

do is guess.

Law students reading this might be nodding along as they have sensed the

inherent ‘squishiness’ (to use a technical term) of the advice their law professors give.

Similarly, associates forced to write for more than one partner may too smile in

agreement, as they have been forced to write in two styles to please two partners – both

of which are sure they know how a good brief is written. Or maybe even the professors

are quietly agreeing, as they have read books like Garner and Scalia’s co-authored book

on writing, Making Your Case: The Art of Persuading Judges, in which the authors openly

disagree on a variety of stylistic choices.

296