Using Credible Incentives to Reduce Smoking in Waymos and Enhance Quality of Life in the Urban "Private Commons"

I’m a big fan of Waymo—the driverless robotic cars that have revolutionized urban mobility. I prefer them over Uber for their reliability and the sheer novelty of cruising without a human at the wheel. But recently, for the second time, I hopped into a Waymo at 8 AM only to be greeted by the unmistakable skunky smell of marijuana smoke lingering from a previous passenger. It’s not just annoying; it’s a classic example of an intertemporal negative externality, where one rider’s actions degrade the experience for the next.

The Ride Share economy creates a “Private Commons” where strangers share private space. It represents an interesting mix of private property and public property that strangers share sequentially.

Waymo operates as a private company, but its vehicles function like a shared commons. We all benefit from the economies of scale that come with ride-sharing: lower prices because the car isn’t dedicated to a single user like a limo with a personal driver. Instead, it’s a sequence of strangers hopping in and out throughout the day. This model relies on an implicit social pact—that each rider leaves the vehicle in a state that doesn’t ruin it for the next person. But as economists know, shared resources often fall victim to the tragedy of the commons, where self-interest leads to overuse or degradation.

The causality here is one-way: A rider at 8:30 AM can affect someone who rode at 9 AM in the same vehicle, but the reverse isn’t true. If an early rider smokes, vomits, or otherwise trashes the car, they impose costs on everyone who follows. Ignoring moral appeals (”be considerate!”), how can we use incentives to foster accountability and protect this shared space?

NOW Waymo does have cameras in the car and it does film the passengers BUT Waymo has stated that its systems are not designed to identify individuals with this data. Clearly, Waymo doesn’t want to be creepy about its surveillance of passengers. Of course, such a monitoring technology would solve the problem that this column explores. But, I proceed under the assumption that Waymo isn’t using these data to pinpoint pot smokers.

Yesterday, after my smelly ride, I immediately contacted Waymo to complain. But complaints alone aren’t enough. What Waymo needs is a system of algorithmic accountability—a smart, data-driven approach to deterrence that doesn’t require invasive cameras (which could raise privacy concerns and costs). Here’s how it could work, drawing on principles from economics and big data.

The Algorithmic Solution: Probabilistic Warnings and Credible Punishments

Waymo already tracks every ride: who gets in, when, and for how long. They have a precise matrix of occupancy for each vehicle, minute by minute. If a rider like me reports an issue—say, marijuana smoke—they can look back at the sequence of previous passengers. Let’s say there were four riders before me that morning. Without camera use, Waymo can’t pinpoint the culprit with certainty, but they don’t need to. Instead, they can apply probabilistic deterrence, inspired by Gary Becker’s model of crime and punishment, where the expected cost of bad behavior discourages it.

The process:

1. Uniform Probability Assessment: Send an email to all four previous riders (or however many are in a reasonable window, like the last hour) stating: “Marijuana smoke was detected in the vehicle shortly after your ride. We assume innocence for now, but if this happens again in a vehicle you’ve used, we may impose penalties, up to banning you from the platform.”

2. Escalating Warnings: Treat it like a “strike” system. The first notice is a gentle nudge—innocent until proven guilty. But if the same user is flagged in multiple incidents (e.g., they’re in the suspect pool for three complaints over a month), the probability of guilt rises, triggering real consequences like temporary suspensions or permanent bans.

3. Credible Commitment: Waymo must publicly announce and stick to this policy. Riders who enjoy smoking in Waymos (perhaps for the “synergies” of a relaxed, autonomous cruise) would think twice, knowing there’s a non-zero chance of getting caught and punished. Even if the odds are low (say, 1 in 4 for a single incident), the threat of losing access to a convenient service acts as a deterrent.

This isn’t about perfection; it’s about shifting incentives. Dope enthusiasts might still take the risk occasionally, but overall, behavior improves because the expected cost outweighs the benefit.

The Benefits: A Win for Riders, Waymo, and Society

Implementing this would enhance Waymo’s product without adding hardware costs. Cleaner vehicles mean happier customers, higher retention, and fewer resources spent on deep cleans . Waymo might lose a sliver of revenue from banning chronic offenders, but the gains in user satisfaction and operational efficiency would more than compensate.

In our big data era, platforms like Waymo have the tools to protect “the commons”. By leveraging ride data, they can establish accountability matrices. This not only reduces negative externalities but also builds trust in the platform, encouraging more people to opt for shared autonomous vehicles over personal cars—potentially cutting urban congestion and emissions.

Broader Implications: Big Data as a Guardian of the Commons

This isn’t just about Waymo; it’s a blueprint for other shared services. Think Airbnb, where hosts deal with messy guests, or public bike shares plagued by vandalism. Algorithmic accountability harnesses data to enforce norms without constant surveillance, preserving privacy while promoting civility.

Of course, there are caveats: False positives could frustrate innocent riders, so the system needs transparency and appeals. And it assumes complaints are genuine—perhaps tie it to verified sensors for smells or messes in future iterations. But starting simple, with probabilistic emails, could yield quick wins.