NEW ENGLAND JOURNAL OF MEDICINE
Augmented Reality Decision Assistance In Teens with Choice-Paralysis Hyperanxiety Disorder
Abstract:
Background:
Teens with choice-paralysis hyperanxiety disorder (CPHD) have historically underperformed in happiness index measures and lifetime earnings compared to their peer group. Augmented reality decision assistance (ARDA) may lead to more optimal decisions in activities of daily living, which compounded over time may ameliorate the detrimental effects of CPHD.
Methods:
We conducted a 48-month, multicenter, double-blind, phase 3 trial involving individuals 18-19 years of age with CPHD who were enrolled in their first semester of university. Participants were randomly assigned in a 1:1 ratio to receive upload of ARDA v3.124 vs. placebo to their augmented reality corneal implants. ARDA v3.124 is a “nudge-based” decision-assistance program trained on a set of 3,898 high-performance individuals. The primary end points were change from baseline at 48 months in score on Alpert Well-Being Self-Fulfillment Index (AWBSF-I; range, 0 to 50, with higher scores indicating greater sense of well-being). Key secondary end points were mean projected salary, social-connection score as measured by Latel Relationship Score Index (LRSI), undergraduate GPA and degree completion rate. A secondary analysis was performed assessing rate of motor-vehicle collisions, suicidal ideation, all-cause hospitalization days, and anxiety adjustment index, in order to replicate beneficial findings previously described in the NO-CRASH, PREVENT-SI, BEWELL, and GARDEN studies.
Results:
A total of 1618 participants were enrolled, with 788 assigned to receive ARDA v3.124 and 830 to receive placebo. Mean AWBSF-I score at 48 months was 42.2 in the intervention group and 31.5 in the control group (difference, 8.7; 95% confidence interval [CI], 8.3 to 9.1; P<0.001). Mean projected salary as measured by Lehman Model was 114,232 USD in the intervention group and 95,878 USD in the control group (difference, 18,354 USD; 95% CI, 12,292 to 24,948). LRSI, undergraduate GPA and degree completion rate, were statistically significantly higher in the intervention group. Previous findings in ARDA trials as listed above were replicated. No deleterious side effects as outlined by the Human-AR Interface Committee (HARIC) were noted. All programming followed strict Asimov Hazard containment protocols, as outlined by Geneva AI Safety Committee.
Conclusion:
ARDA v3.124 significantly improved well-being in individuals with CPHD as measured by a universally-validated measurement index, and also resulted in higher earning potential, greater social connection, and higher academic performance as measured by GPA and graduation rate. We estimate the improvement in well-being index is equivalent in magnitude to a windfall of ~73,000 USD per year. Further research is needed as to whether these beneficial effects translate to the general population.
Late one night, when the rest of the staff had left for the day, Hilary slipped back into the containment room and downloaded ARDA v3.124 into her own corneal implant.
She was careful to cover her tracks. She used a corner of the office that she knew had low visibility from the security cameras. She used a port-connection cable with a ID filter so that it couldn’t be traced to her implant. And she registered the download under a dummy trial participant’s name, which she then excluded from the searchable registry.
Pulse pounding, she left the office through the backdoor, and waited until she was home before opening the zip-file.
She had helped design the delivery mechanism, and so was familiar with the installation process. Even then, it was clunky - very clearly a clinical trial program rather than one optimized for commercial use.
Hilary had had her corneal implants for more than a decade. Like most people, she used them as a replacement for video monitors, and for navigation when driving.
She didn’t know what she was expecting, but when her implants rebooted, the world felt unchanged.
“Siri, turn on ARDA,” she said out loud - and was quickly informed that ARDA was already installed and running. She walked a few circles around her house, but there were no nudges, no prompts - nothing. The calibration process was only supposed to take a few minutes, but there were a few outlier participants for whom it took a whole day. She wondered if she was one of them.
The next morning, it was a Saturday, and for once she didn’t sleep in. After a few minutes she realized what it was - the light coming through the bedroom curtains was brighter than it should be. It looked gorgeous - like the hand of God sunbeaming into the room.
She went to her closet, and at first she thought that her clothes were missing. But then - no - it was that ARDA had muted the color of most of her clothing into the background wall. Only two pieces of clothing seemed to really stand out - a shirt and pair of old leggings that were what she would have chosen anyway.
With a flicker of her gaze, she clicked on the ARDA settings in her corneal implants:
Version number: 3.124
Nudge strength: 4/10 (recommended default)
Field of vision: Full (recommended default)
Intrusivity: 3/10
Mutability: 3/10
Share data: Off
Color drift: 0% orthogonality
Going to the grocery store was a revelation. Entire portions of the shelving were muted, and only one or two items for each category of item stood out. Notably, in the cereal aisle, the sugary options had a gray tinge to them, looking drab and unappealing. By the cash register, ARDA had replaced the candy sections with panels of empty shelving that blended seamlessly into the existing shelves. The only reason she knew it had changed was because she had bought a Snickers bar from the shelf the week prior.
Even though she had known what to expect, even though she had helped design some of the features themselves, she was still surprised the first time a true “nudge” happened.
She had been driving home, when a section of the road to her right became brighter, more vivid, and somehow more beautiful, tugging obviously at her attention. She turned the corner, and followed the patch of road another few blocks, where it ended in a parking spot.
It was a street jazz festival. She had heard that it had been happening, had meant to go, and then completely forgotten. She checked the local location sharing board, and saw that a few acquaintances from college were gathered by the MacMulligan’s where two trombonists were playing an improvised duet.
She hesitated. She hadn’t been planning on talking to anyone today. She had gone to the grocery store originally with the intention of picking up some ice cream and watching TV. for most of the day, with a margarita or three in the afternoon.
But she was already here… and so she entered the crowd, heading towards MacMulligan’s.
Moving through a crowded space felt less claustrophobic than usual. She realized why - ARDA was visually thinning out the crowd in the distance, adjusting the person-density to something less overwhelming. She had always hated crowds, and now it had almost a pleasant feel to it, an unobtrusive gentle bustle.
When she approached the group of acquaintances already gathered at the bar, she felt a familiar fluttering in her chest. It was the same feeling she had when approaching strangers, or before giving a talk in front of an audience. She paused for a second. She could still go home, she thought to herself. It was only a ten minute drive away. There was no reason she had to be here.
But then, this was the entire reason she’d downloaded ARDA. She continued.
ARDA charted a path through the crowd - and she realized it had course-corrected her, so that she was approaching them from the front rather than from behind.
One of them - John, his name was John she remembered - saw her and waved her on over.
Ah, she thought. So much less awkward than approaching from behind and tapping on someone’s shoulder like a creep.
She said hi to the group of four sitting at the bar, and ordered a drink to calm her nerves. John’s face seemed open and friendly - was that really his smile? Was ARDA making his face seem more friendly than normal? - but she found herself struggling to make conversation. She was reminded once again why this wasn’t her forte.
It was 5 minutes and 4 awkward pauses into the conversation that she felt another nudge. The settings icon in the top right of her corneal implant screen started glowing like a tiny second sun. She clicked it, and the ARDA settings appeared.
Nudge strength: 4/10 (recommended default) was highlighted on the menu screen. She clicked it, and on the slider, the number 8 glowed gently…
She hesitated, and then made the change.
Nudge strength: 8/10 (maximum level available for clinical trial participants)
She realized that someone was saying her name.
“Hilary? You alright?” one of the women was saying from across the table. Hilary couldn’t remember her name.
Friendly, non-obtrusive words appeared above the woman’s head.
Name: Janet Hawthorne
1: *1.2 second laugh* Sorry, just wool-gathering. Your brain ever just hit you with an embarrassing memory out of nowhere?
2: Ah man. That bartender is distractingly cute. Sorry, what did you say?
With startling ease, Hilary read off the first option. Somehow, the words came smoothly, like they had always existed within her. Or at least, within a bolder, more graceful version of herself.
“Oh all the time. Sometimes I feel like I’m arm-wrestling with my own brain,” Janet replied.
ARDA’s response was there before Janet had even started her sentence: Amen, I’ll drink to that. *Raise glass*
“Amen, I’ll drink to that,” Hilary said, raising her glass.
To her surprise, everyone around the table raised their glasses and all took a drink. She waited for ARDA to provide her a prompt, but none came.
She thought of something.
“Say, Janet. There’s something I’ve always wanted to ask you, ever since we graduated,” Hilary began.
Somehow, the rest of the afternoon played out like a dream. ARDA gave prompts only a handful more times, each time before she herself realized she was going to be awkward. The sheer knowledge that ARDA was waiting in the wings, ready to step in as a safety net, gave Hilary an exhilarating confidence.
When the group finally dispersed, it was reluctantly, and with concrete plans to hang again the next weekend. For the first time in her life, Hilary felt like she was part of the in-group.
In the weeks afterward, every moment with ARDA felt like an affirmation that she had made the right decision.
Part of it was that some choices were made for her and she could just follow ARDA’s railroading: from the cereal she bought, to her retirement plan options, to some of the coding decisions she was making at work as they worked on ARDA v.3125.
Part of it was that ARDA nudged her towards certain social gatherings. Suddenly, she had a regular friend group that met up every week. She became known not quite as the life of the party, but as the heart of the friend group, the lynchpin that connected people and pulled them together.
Part of it was that ARDA made certain visual alterations to her world. Sunsets were enhanced and more beautiful. Billboards she found distracting were simply muted out. She suspected that the facial expressions of her coworkers were being altered to decrease her social anxiety and make her more confident at work.
One day, she realized that ARDA had been muting out all the homeless people she was used to seeing on the way back from work. She hadn’t realized how much moral distress they had been provoking in her, how many microscopic adrenaline spikes she’d felt at just their presence. Now, the homeless were not quite invisible so much as blended seamlessly into the background.
It was the neighbor’s car that finally made her feel doubt. A brand new Mercedes, more beautiful and expensive than she could have conceived of affording. She had passed by it without noticing for what must have been weeks, since ARDA had not only muted it but changed its form so that from a distance, it looked like a cheaper brand. She’d only noticed because she ran into her neighbor, who casually name-dropped it into the conversation.
She felt a flash of envy - and then the emotion was replaced by doubt. What else had ARDA been muting or shapeshifting without her even noticing? She’d known that there was a department which had been focused on decreasing interpersonal comparison with more well-off peers, as part of the satisfaction-optimization thrust. But she’d never been closely involved with their projects. And it was one thing to hear about it in passing in the abstract, and another to experience it viscerally. Was it changing parts of her vision whole-sale? Were there objects that were not just muted but made literally invisible? Were there blindspots within her blindspots she’d never be able to see, simply because there were no visual loose threads to pull on? What were the unknown unknowns?
She went to her settings.
Nudge strength: 8/10 (maximum level available for clinical trial participants)
She thought about moving the slider back down to its factory settings. All it would take was one click. It would be easy. Beneath the surface of her consciousness, a flicker of attention noticed that the numbers below 8 were grayer than before, the font altered to make them less attractive - and then the thought passed.
She thought about the past few months. Her life had been improving, and those improvements seemed to be demonstrating compounding interest. After a series of tiny nudges, she’d been able to positioned herself for a promotion to be technical lead of the ARDA infra team. It came with a significant pay raise, and it had always been something she’d never tried for due to fear of failure.
She thought about Ben, who she’d gone on four dates with, and who seemed to be more into her than any of her previous exes. Who she could talk to confidently, because she knew that ARDA was there to provide conversational bumper rails if she found herself at a loss.
She thought her group of friends she could call on at any time, who she could tell anything to. She never felt awkward at parties anymore, even when she didn’t know anyone there. She never felt lonely in a room full of people anymore - it didn’t feel like she was popular so much as that she belonged.
She thought about her relationship with her sister, which was the best it’d been in years. About the 10 pounds of weight she’d lost without even trying, presumably because she was being nudged towards less junk food. About the fact that she’d filed her taxes ahead of time, which would have been previously unimaginable.
She looked at the slider, which was currently set at 8 out of 10, the maximum allowed setting, and in the end, the decision was the easiest she’d ever made.
“Let’s jailbreak this thing,” ARDA-Hilary said.
Hee! Excellent. Creepy!
I was going to write something like "Whoever designed the feature to allow the software to suggest adjusting its nudge level is really to blame", or "No, it's whoever allowed it to suggest technical changes to software including itself"... But of course at least the second one of those wouldn't have been a deliberate choice by any human, just a consequence of the neutral network training...
Wow, I want this please?
Sign me up!
You can have a share of my new income.