The Slippery Slope of Technology Design

Technological change rarely announces itself with a sudden, catastrophic lurch. Instead, it moves gradually, descending a gentle gradient paved with good intentions, convenience, and incremental optimization. We often dismiss the “slippery slope”—a paranoid reflex assuming that any minor shift in the status quo will inevitably send us hurtling toward dystopia. But when analyzing technological evolution, the slippery slope is not an illusion; it is a predictable trajectory governed by two powerful forces: affordances and incentives.

An affordance is a technology’s inherent capability—what it makes something naturally easy to do. Gunpowder in a metal tube affords poking holes in things from a distance. While initially used for weaponry, it also naturally afforded the creation of powder-actuated tools for driving nails into concrete.

Incentives, on the other hand, are the economic or social drivers that push us to explore those new affordances. When a technology makes an action faster, cheaper, or more profitable, the temptation to expand its use becomes irresistible, often pushing the original, noble intent into morally murky territory. And, thus, the slide down the slippery slope begins.

The Vision Machine and the Evaporation of Obscurity

In 2018, while driving south on Interstate 80 for Thanksgiving, Brian Hofer and his brother were pulled over by police. It was not a routine traffic stop; officers drew their weapons, forced Hofer’s handcuffed brother to his knees on the asphalt, and pointed a gun at his head. Forty terrifying minutes later, the police released them, realizing their mistake.

The culprit was an Automatic License Plate Reader (ALPR). An optical character recognition camera had scanned Hofer’s rental car and matched it to an outdated “hot list” of stolen vehicles, instantly pinging law enforcement.

ALPRs were originally designed with a narrow, highly beneficial intent: to track stolen vehicles or locate suspects in real-time within a specific region. But the technology’s affordance—the ability to scan, log, and cross-reference every passing license plate—created an incentive for mission creep. Instead of merely checking against a specific hotlist, modern ALPRs capture and store data on every passing car, creating vast databases of cars over time and place.

This shift transforms what was a tool for finding stolen cars into an infrastructure for mass surveillance. By early 2025, a suburban Chicago police department was asked by Texas law enforcement to supply vehicle location data to track a woman seeking abortion support in Illinois. This request tapped into a network of 83,000 cameras operated by Flock Safety, a private company. Police departments and the Customs and Border Patrol have also used these systems to monitor citizens legally attending protests, such as the fall 2025 “No Kings” demonstrations. [Cox]

A similar, perhaps steeper slope can be observed in the evolution of facial recognition. In its infancy, facial recognition was confined to military and intelligence applications, such as DARPA’s FERET program, or tightly regulated environments like DMVs to prevent identity fraud. The mission was narrow: one-to-one verification of a known subject.

Yet, as computer vision algorithms improved, the affordance expanded from identifying known criminals to identifying everyone.

In 2022, Kelly Conlon, a mother chaperoning a Girl Scout trip to Radio City Music Hall, was barred from entry to see the Christmas pageant. Without human intervention, a facial recognition system flagged her as a prohibited person simply because she worked for a New Jersey law firm engaged in litigation against the venue’s parent company, Madison Square Garden Entertainment. And it turns out that the company, MSGE, had a set up a policy to exclude people from law firms engaged in action against them. [Wallace]

Today, third-party firms like Clearview AI scrape billions of non-consensual images from the public internet, boasting a 99% accuracy rate to attach a name to almost any face. Even grocery chains like Wegmans deploy AI to identify individuals deemed a “risk.” [Hill]

This widespread application obliterates what legal scholars term “practical obscurity”—our reasonable expectation of anonymity in public spaces. When a system like New Orleans’ privately-run Project NOLA scans every face in the French Quarter without the oversight of public records laws, we have fully descended the slope: watching and identification has become the default.

We see this in product announcements like Meta’s “super sensing” glasses could continuously record audio and frequently snap photos. Interestingly, Meta tells us that the “raw footage and audio would not be stored by Meta or made available to the user…” Who, then, is the data being recorded for? [Peters]

The Panopticon We Built Ourselves

The proliferation of surveillance was not violently imposed upon us; it crept in through convenience. In the 1970s, a closed-circuit television camera in a bank lobby was a rare, expensive piece of hardware that signaled security. Today, the lenses are everywhere, and we are the ones installing them.

The slippery slope flattened out into a plateau of mutual observation because cameras became decentralized. We welcomed doorbell cameras to catch package thieves. We bought electric vehicles, like Teslas, that continuously record their 360-degree surroundings for insurance claims, inadvertently creating mobile surveillance networks that police routinely requisition for evidence. We carry pocket-sized supercomputers that effortlessly record high-definition video and track our movements throughout the day by GPS.

From above, cheap drones and aerial cameras operated by firms like Persistent Surveillance Systems can capture a five-mile radius, turning people into trackable, moving blurs. The moral friction once associated with surveillance has vanished. We have all become participants in a networked ecology of eyes, driven by the affordance of cheap digital storage and the incentive of personal safety.

The Economics of Enshittification

The mechanics of the slippery slope extend far beyond hardware and surveillance; they dictate the digital economies we inhabit. In the early 2000s, software was a straightforward transaction: you paid for an application once and owned it forever.

But in 2011 and 2012, Apple and Google introduced support for in-app subscriptions, fundamentally altering the developer incentive structure. Instead of a one-time purchase, companies realized they could generate a continuous stream of recurring revenue by locking features behind monthly fees. This model exploded; the global in-app purchase market is projected to skyrocket from $171.7 billion in 2023 to nearly $688.0 billion by 2033.

This economic shift birthed a phenomenon technology writer Cory Doctorow famously calls enshittification. Doctorow outlines this descent as a three-step process, a kind of embodiment of the slippery slope:

• First, companies offer a fantastic product, often operating at a loss, to attract and lock in a critical mass of users.

• Second, once users are trapped by network effects or a lack of alternatives, the platform begins to abuse them, shifting value to corporate customers like advertisers or publishers.

• Finally, the platform claws back all remaining value for itself, leaving the users with a degraded, ad-inundated experience where basic functions require constant payment.

The affordance of wireless software updates for everything has allowed this model to bleed into the physical world.

In 2020, BMW piloted a program where drivers had to pay a subscription fee to unlock heated seats—hardware already physically installed in the car. While U.S. consumer backlash forced BMW to retreat temporarily, the model thrives elsewhere; in South Korea, permanent access to Apple CarPlay costs over $304, and a heated steering wheel runs roughly $10 a month. As cars increasingly become software platforms on wheels, the incentive to disaggregate services and charge for every component will only grow.

The in-app purchase model allows a slippery-slope slide towards a world where everyone has to pay to play for everything.

Good Intentions, Toxic Outcomes: The Juul Paradox Story

Perhaps the most tragic illustration of a well-intentioned innovation sliding into disaster is the genesis of modern vaping. The original mission was profoundly noble: to eliminate the lethal harms of combustible cigarettes.

Stanford industrial design students James Monsees and Adam Bowen, both smokers, wanted to solve a global health crisis. They aimed to turn tobacco into a “luxury good” rather than a deadly drug delivery device. Their early prototypes, however, failed a crucial affordance test: they could not deliver nicotine efficiently enough to satisfy heavily addicted smokers. [Dunworth]

To solve this, they turned to secret historical research from R.J. Reynolds and hired a chemist to mix pure nicotine with acid, creating “nicotine salts”. This innovation eliminated the harshness of high-concentration nicotine, allowing for a massive, smooth hit. Paired with a sleek, USB-like design, the device—Juul—was born in 2015.

The system worked entirely too well. The affordance of smooth, highly potent nicotine delivery, combined with the incentive to capture a massive market with flavors like “Watermelon Ice,” sparked a public health disaster. Instead of just helping adults quit smoking, Juul ignited a youth epidemic. By 2017, one in five high schoolers was vaping, consuming cartridges that contained the nicotine equivalent of an entire pack of cigarettes. In 2018, the FDA declared youth vaping a “national epidemic”.

A technology designed to save lives ended up worsening the problem for a new generation of never-smokers, creating profound addictions and introducing the threat of a “gateway effect” back to combustible cigarettes.

Navigating the Descent

The history of technology is littered with tools that escaped their creators’ original intent. A system designed to find stolen cars becomes a dragnet for abortion seekers; a program to catch identity thieves bars a mother from a Christmas show; a model to support app developers turns our cars into ransom-ware; a device to cure smoking addict’s teenagers.

The lesson for us here is not to halt innovation, but to recognize that the slippery slope is a structural feature of technological progress, not a logical fallacy.

When we introduce new affordances into the world, the incentives of the marketplace will invariably push those capabilities to their absolute limits. To build responsibly, we must look beyond the immediate problem we are trying to solve and anticipate the adjacent possible—because once a technology makes something easy, someone, somewhere, will inevitably do it.

==============================

References

Cox, Joseph; Koebler, Jason. (May 29, 2025) “A Texas Cop Searched License Plate Cameras Nationwide for a Woman Who Got an Abortion” 404 Media. https://www.404media.co/a-texas-cop-searched-license-plate-cameras-nationwide-for-a-woman-who-got-an-abortion/

Dunworth, James. (March 30, 2023) “The True Origins of Vaping (It All Started in the 1970’s!)” E-Cigarette Direct (website) https://www.ecigarettedirect.co.uk/blogs/ashtray-blog/vaping-origins-1970s

Hill, Kashmir. Your Face Belongs to Us: A tale of AI, a secretive startup, and the end of privacy. Penguin Random House. (2023)

Peters, Jay. “Meta is reportedly working on smart glasses that would be recording all the time” The Verge (July 8, 2026) https://www.theverge.com/tech/963138/meta-smart-glasses-recording-super-sensing-ai

Wallace, Sarah. “Face Recognition Tech Gets Girl Scout Mom Booted From Rockettes Show — Due to Where She Works” (Dec, 2022) NBC News, Channel 4. https://www.nbcnewyork.com/investigations/face-recognition-tech-gets-girl-scout-mom-booted-from-rockettes-show-due-to-her-employer/4004677/