“Anything really rich?”
He means “rich with data,” like a vein of ore. “We got lucky with a few private adult chat profiles. Lots of video.”
He raises an eyebrow. “No problems with using those?”
I shrug. “The prohibition is against ex parte communication. As far as I’m concerned, if you reuse credentials already leaked in another data breach, you’re just asking for the world to stroll in and take a look-see.”
He nods and forks off fifty new casts—blank neural networks to be populated and trained with the data gathered by the harvesters. He types into a bunch of console windows and flips through colorful displays, visualizing the developing idols. Even with the full processing power of the modeling room at his disposal, it still takes time to build low-res idols for the voir dire stage.
I’m a little bored and feeling restless. I don’t mean to sound boastful, but there really isn’t anyone in the city—maybe the whole east coast—better at this than I am. The work has gotten a bit stale.
I decide to call home.
“Hey.”
“Hey. Am I going to see you this weekend at all?”
I melt a little at Dylan’s voice, yearning, with just a hint of desperation in it.
“I don’t think so. I told you the trial starts on Monday. I promise I’ll make it up to you Monday night.”
As I talk to him, my eyes roam around the modeling room. Some of the analysts are tapping away at their keyboards, assisting Kevin; others nap in their cubicles, knowing that they’ll be asked to pull an all-nighter. Along the eastern wall are storage cubicles for the embodiment robots, into which we could pour the idols of the various trial judges that attorneys at the firm regularly argue in front of. Two technicians are doing some maintenance on the robot that currently hosts the idol for Judge May. Tomorrow morning, starting at 8:00 AM, the trial team will come in and hold mock sessions all day in front of her. I’ll need multiple simulated juries of idols ready for them.
“I suspect you’ll just drive back to the office right after I fall asleep,” Dylan says. A loving chuckle.
I laugh too. He accepts how much I love my work, even though he finds it hard to understand. We accept quirks in the people we love.
“You should take me up on visiting the office sometime,” I tell him.
He makes some noncommittal noise. I know he finds the idea of litigation idols creepy. But what fun thing isn’t at least a little bit creepy? Dolls, Imojis, those Furby-things I had as a kid.
“I talked to my dad tonight,” he says. “Wanted to find out when he first knew he wanted kids.”
Case in point. I don’t understand why he talks to an idol of a man who’s his father only in name. But I accept it.
My mind drifts as he recounts the conversation.
Along the north wall, sitting in their black robes, are the eleven active and senior judges of the circuit court of appeals. The firm actually hired a local artist to sculpt the molds for the faces. Cast from the latest in biomimetic materials, the robot faces faithfully replicate every fraction of a millimeter in a raised eyebrow and every subtle wrinkle at the corner of the mouth in an exasperated sigh.
“...‘It’s all too easy to become obsessed with yourself,’ he said...”
The appellate judge-bots are serious overkill. We argue less than a dozen cases at the court of appeals every year, and oral argument is the least significant stage of an appellate case. Why can’t the partners just practice against the idols in VR or rely on plain video? The realism of these robots adds virtually no advantage. Personally, I feel the money would be much better spent on upping the simulation resolution in the trial judge idols.
“...‘Family’s important, you know?’ I know he wasn’t really my family, but I feel the same way...”
But that’s not my decision. The prestige of appellate litigation makes the managing committee go weak at the knees and happy to throw money at it. The partners like to practice in front of the robots and impress important clients with private tours of the modeling room.
(Mind you, I’m not saying analytics aren’t useful at the appellate stage. Judges are people, and it is possible to craft briefs to appeal to their peculiar tendencies. Even better, the junior associates often can give us private information about the judges’ clerks—their former classmates—and if your brief can get the mini-idols of the clerks excited, chances are, you’ve got yourself a free advocate in the chambers. To keep the appellate idols up to date, Kevin and his team download every appellate record, feed the transcripts and recordings to the idols, and test the predictions against actual outcomes. I understand that so far they are batting just over .900, which I have to admit is pretty damn impressive.)
“...What do you think?”
Too late, I realize that he’s expecting an answer of some sort. “I’m... I’m sorry. What?”
Dylan sighs. I can hear the disappointment as well as the forgiveness. He knows I haven’t been listening. “I was suggesting that we take a road trip together. You have the vacation days. No phone, no tablet, no idols, no work at all. Just driving and talking. Just us.”
“...What brought this on?”
“I think we need to talk about the future, about kids.” He sounds calm, but isn’t.
I feel blindsided. Why can’t he just keep up his end of the comfortable routine, the routine I rather like? What is this sudden emotional outburst about? There’s been no warning.
“I... Things are really busy. I can’t think about—”
“Okay.” Kevin spins around in his chair. “The roughcasts are ready. You want to sculpt?”
“Gotta go,” I say into the phone.
“Love you,” he says after a pause, hurt, still yearning.
“Love you too,” I say, and I mean it. I hang up.
I take a deep breath and exhale, trying to push away the aftershock of the conversation with Dylan. I can’t think about this right now. I have to win first.
I scoot my chair near Kevin’s giant monitors, filled with a grid of rotating, amorphous, rainbow-hued blobs. The visualization software is showing various personality traits of the crude idols, and it’s my job now to tweak them based on my intuition about the prospective jurors.
Sculpting idols is part science, part art. You know how those wax portraits done by 3D scanners sometimes don’t seem to capture the “soul” of the subject as swell as a bust crafted by a good artist? It’s the same principle. You need the human touch.
I click on the first square in the grid until the spinning blob expands to fill the screen. I bring up the scraped data on prospective juror number one, and begin to examine the idol with the mouse. I tell Kevin where I think the algorithms didn’t quite get it right, and he modifies the model per my instructions.
IT’S NEAR MIDNIGHT by the time I bring the junior associates the news that the idols are ready to confront them.
“The software has already ranked the jurors in terms of how desirable they are to us. If your own assessment is different, note that somewhere, but don’t question the algorithm! The machine never misses anything. Figure out what you missed.”
Before we had idols, consultants used to conduct community surveys and neighborhood roundtables before a trial to advise the attorneys on general attitudes in the jury pool. And then there would be a mad rush to segment the potential jurors by demographics, profession, tax bracket, location, and the like right before voir dire. Compared with those crude tools, even the low-res idols we could construct overnight are precision scalpels.
“Your job is to figure out lines of questioning that will allow us to strike the ones we don’t like for cause—or better yet, to get the other side to strike them for cause or waste one of their peremptories. Get them to voice their prejudices, conspiracy theories, wacky weltanschauung. And if there’s time, also figure out how we might save the ones we like with good rehabilitation questions. The software will give you suggestions, but you need to vet them and see if you
can come up with a plausible script of questions that won’t be too obvious and tick off the judge. This is where you can prove you’re better than a machine!”
Never hurts to try to encourage the troops a little.
I watch the junior associates run back to their offices with their assigned idols to probe and prod, feeling like a wise Jedi master sending her Padawans into battle.
They’ll do fine. I’m not saying voir dire research is easy, but working with the crude idols roughed out with so little research and time isn’t too challenging. The machine’s suggestions for striking undesirable jurors are almost always good enough. The truth is: the other side will have idols of their own and will be prepping just as hard, and they’ll never allow potential jurors with a pronounced bias towards us to be empaneled. We’ll end up with a jury that’s reasonably persuadable either way. When I explained this to Dylan, he looked horrified. But I told him this is just the system working the way it’s supposed to, assuming you think a bunch of fence-sitters swaying whichever way the hot air blows is the best way to achieve justice.
“You tell me you like your work. But you sound so cynical about it.”
I don’t know why I’m thinking of Dylan’s words. They bother me more than I care to admit. But there’s no time for that.
I turn to the much tougher task: opposing counsel and witness prep.
I load up the idol of the lead partner on the other side, a woman who has been litigating these types of cases longer than I’ve been alive.
She stares out of the screen, grim-faced, lips pressed together severely. I can see how she could put the fear of God into a first-year associate with one look. But she doesn’t intimidate me. With so much trial experience in the public record and speeches at professional conferences, we have a lot more transcripts and videos to work with, lots more data to feed into her idol.
“How do I make you mad?” I whisper into the screen.
She remains frozen, unable to respond.
We’ve been at this for days, weeks. The daily battle of wills has become a part of my routine, like filling out time sheets and doing the dishes. I’ve found several openings already, though none quite at the level of a killing blow. Not yet.
The software scours the ether nonstop for information about her, relentlessly refines her idol. I’m going to try again tonight.
I press the button to make her come alive.
The idol has no memory. Every day is brand new. I begin, as I always do. “Good evening, Ms. Gaughen.”
She looks irritated. “Do I know you? Do you have an appointment?”
Though she’s done this to me countless times already, I suppress a pang of... what? Annoyance? Injured pride? Of course she wouldn’t know me. Being unknown, devoid of prestige, no matter how much I contribute to a victory, is part of the job. On the firm website, I’m listed in the “Tax and Private Client” group. Good place to stay out of the limelight.
“I’ve admired your work for a long time, Ms. Gaughen. I’m feeling a little stuck, and I’d like some advice on my career.” I have all the opposing counsels’ idols primed to respond to this question. It’s a bit artificial, but it’s the quickest way to get to business.
“All right.” Her face relaxes. “Tell me about yourself.”
My first impulse is to lash out. The question, one I’ve heard so many times from her before, feels like an accusation tonight. I don’t understand. What is wrong with me? What is this emotional outburst? There’s been no warning.
I force myself to follow the script.
In litigation, one soon learns that whether something is true doesn’t matter; whether the jury believes it’s true is the only thing that matters. This isn’t a criticism. We designed the system that way. Since the jury cannot conduct experiments, examine witnesses themselves, or investigate the evidence independently, they’re solely allowed to decide who to believe. Credibility, authority, truthfulness—we make these assessments based primarily on hunches and emotion, and that makes them open to manipulation. Some of these courtroom manipulations have ancient roots: the way a lawyer dresses, the vocabulary she uses to talk to the jury, the list of impressive acronyms and institution affiliations deployed to prop up an expert witness. The law has developed techniques for confining the scope of these tricks.
But idols allow other manipulations, ones that the law hasn’t quite caught up to.
While I’m giving Margaret T. Gaughen, Esq., my fabricated resume (to get realistic responses out of the idols, the questions have to be as lifelike as possible, hence the charade), the idol-probing software pops up a line of suggested irritant questions that I haven’t seen before on a screen to the side: “She seems to have a higher-than-usual hostility towards law review alums.”
I frown. Really?
The software shows me a transcript of a deposition from years ago, highlighting one exchange in yellow.
Witness: I was on the Law Journal. I know what I’m talking about.
Counsel: If I have a question about the Bluebook, I’ll be sure to ask you. Stick to the facts. You’re not an advocate right now.
Next, the software brings up a video clip, a Q-and-A session after a lecture at a law school. A student raises her hand to ask about the most important quality Gaughen looks for in an associate.
I don’t care if you have straight A’s or if you’re on the law review. Actually, it would be better if you weren’t. Then at least I’d know there’s a chance you don’t already think you know everything.
I confirm that Gaughen hadn’t been on the Law Review at HLS, so maybe there’s some lingering sting of rejection. But it seems weak sauce to me. Criticism of the obsession with the prestige of law review membership is commonplace in the profession. This hardly feels like the way to get Gaughen to lose it.
Still, it’s worth trying.
“I was Notes Editor on the Law Review,” I tell the idol in the screen, infusing my voice with a hint of humble pride. “I really treasure the memories.”
I can see her lips purse in distaste. Maybe the software is onto something.
One of the surest ways for a lawyer or witness to lose credibility in the eyes of the jury is to appear to lose control, to lash out in an enraged outburst. Every human being has emotional weaknesses that can be exploited, buttons that can be pushed. Skillful litigators in the past relied on instinct as they probed for such openings during argument or cross-examination, hoping to stumble on a tender spot.
With idols, the search for such openings can be systemized and made a hundred times more effective. From years of trial records and voluminous depositions, it’s possible to construct very high-resolution idols of key witnesses and opposing counsel. The software and I would then key in on exploitable triggers.
“Do you think I should reach out to my old friends?” I ask innocently. “Maybe focus on friends who were also on the Law Review?”
The idol’s face turns even more stony. It really does feel like I’m touching a nerve.
You’d be surprised at the kinds of things that push people over the edge. Once, for example, I got opposing counsel, a litigator with decades of courtroom experience, to scream at us, mouth foaming and arms flailing, in response to a suggestion about taking an early lunch. Sitting as an anonymous observer in the public benches, I could see the shock in the eyes of the judge and the jurors as they watched the bailiff rush over to restrain the man. We settled that afternoon very favorably for my client.
What the jury and the judge didn’t know was that I had instructed our team to consistently adopt a series of mannerisms and idiosyncratic pronunciations that evoked memories of the opposing counsel’s deceased father. You know how, even as an adult, an innocuous word said a certain way by a parent can still instantly revert you both to patterns established when you were thirteen? It was like an extreme version of that. The opposing counsel and his father had had an extremely negative relationship, possibly abusive, and the idol simulations showed me that if we kept it up,
he would eventually break down in front of the jury.
“I suggest that you stop thinking about your time on the Law Review as some kind of proof of your brilliance,” Gaughen’s idol says to me bluntly. “No one cares about that. What have you done?”
Injured, burning pride swells in my chest. I want to give her a litany of my unsung victories, the briefs with my idol-driven insights but not my name, the settlements driven by my idol-tested scripts—I swallow the impulse. I am really distracted tonight. This isn’t about me. I’m not even talking to a real person.
On the side display, the wavering lines charting her simulated heartbeat and blood pressure are way up. I can feel my own heart pounding and my face growing flushed. I take deep breaths and force myself to calm down. It does seem promising. Now, the question is how do I construct a plausible script during trial to push her on this...
Wait. Something feels oddly off about all this.
I pause the idol and turn my full attention to the probing software. Why hasn’t her sensitivity towards the perceived prestige of law review membership come up before?
A few key strokes later, I have the answer. A couple of days ago, one of the gossip blawgs posted a call for stories of partners behaving outrageously, an evergreen topic for these rags. An anonymous poster added a comment saying that a partner at their firm would make all junior associates redo the same research memo over and over instead of giving real training. When another commenter expressed skepticism that a partner would waste firm resources in this way, the first poster admitted that it happened only to them, and it was because the partner wanted to “put me in my place, ’cause I was on law review and she wasn’t.” The software had scanned the commenter’s posting history to de-anonymize it and traced it to Gaughen’s firm, and then, based on other clues in the comments, determined that the poster was talking about Gaughen. The addition of this one piece of data pushed the probability of her idol overreacting to this line of irritant over the threshold.
Do I believe the machine?
Made to Order Page 13