Every week, Americans take over half a million rides in autonomous vehicles across a dozen major cities. DC isn't one of them – but a new bill from Councilmember Charles Allen could change that.

Allen, along with Councilmembers Brooke Pinto and Matt Frumin, introduced the Autonomous Vehicle Deployment Authorization Amendment Act of 2026 on May 1. I've spent a decade in AV policy and, for better or worse, this is the most comprehensive and complex AV deployment bill I've seen from any U.S. city or state. It creates a real pathway to commercial AV rides in the District, while seeking to tackle a range of concerns surrounding AVs including congestion, equity, transit, labor, and liability.

The introduced text is an encouraging start – but it will all come down to how the bill is shaped by the Council.

Autonomous urbanism is a broader vision for a future where AVs unlock tremendous improvements in city life by complementing public transit and reducing car dependence, making neighborhood streets safer for everyone (especially pedestrians, cyclists, and children), enhancing mobility for under-served communities, and supporting the local economy.¹

As we discussed at The Innovation Majority’s April 8th event, “Why not DC? Unlocking AVs in the District,” AVs present a significant opportunity for the District to deliver on safer streets, address inequitable access, and diversify and revitalize the local economy.

Three policy elements give me hope that the bill could set the foundation for DC to become a national leader in autonomous urbanism:

The bill ambitiously seeks to address every type of fear about AVs – but, in the process, treats every potential negative impact as a real threat rather than a distant possibility, from displacing driver jobs to exacerbating congestion to gutting public transit. Some aspects of this bill could lead to more bureaucratic busywork, redundant commissioned reports, and premature actions. These are mostly benign, generally forgivable, and on-brand for DC – but a few provisions create new issues in the service of solving hypothetical problems. 

Three policy mechanisms could be cures that are worse than the disease they propose to treat:

I hope that Allen and his colleagues will find a way to balance their concerns about the hypothetical harms of AVs with the tangible benefits they’ll deliver for residents – and I’ll share a lot more on the topic of autonomous urbanism in the coming weeks and months.

In the meantime, here’s a detailed look at the key policy mechanisms in the bill: what works, what’s unproven, and opportunities for improvement.

DC passed one of America's first AV testing laws back in 2012, alongside Nevada, Florida, and California. Those other states continued to move forward in the past 14 years by overseeing testing, writing new rules, and eventually authorizing driverless rides. They’re now national leaders in AV deployment and policy development, representing a substantial share of the hundreds of thousands of driverless rides Americans take every week in AVs.

Meanwhile, DC stalled. DDOT was tasked with developing a driverless testing permit and delivering a report to the DC Council with recommendations for regulating AV deployment – but, for any number of political or budgetary or bureaucratic reasons, neither have materialized as a final product.

Allen's bill seeks to break from that pattern by setting firm and specific deadlines: a driverless testing permit within 60 days of enactment, a commercial AV permit within 180 days, and establishing a concrete authorization date for on-demand ride networks beginning January 1, 2028. 

It’s a welcome departure from the "wait for consultant reports" and “form working groups” approach that has defined DC's AV posture for over a decade. Additionally, the ambitions of this program will be met with dedicated funding from permitting and VMT fees, which can help address the staffing and resource constraints faced by the responsible teams at DDOT.

The bill creates a Commercial AV Program for DDOT to evaluate applications, issue permits, and oversee deployment. 

To qualify, an operator must have tested in DC for at least 180 days with over 250,000 miles driven, submit a first responder interaction plan (including motorcades and public demonstrations, because DC), a continuity of operations plan for outages and extreme weather, and carry $5 million in liability coverage. 

Application fees are $1 million, and the initial permit fees are a whopping $5 million.

Fleet size is capped at 200 vehicles until the operator submits a comprehensive plan addressing parking, deadheading reduction, and equitable access across all eight wards. After that, DDOT can set higher fleet limits through rulemaking. Ramping slowly is understandable – but, as discussed above, policymakers should be mindful that caps that linger too long can enshrine inequitable access rather than prevent it, and worsen other issues they purportedly seek to address.

To many of those concerned about transportation funding in America, replacing the gas tax with a vehicle miles traveled (VMT) fee is seen as the holy grail that is always, for some reason or another, just barely out of practical reach. 

Allen’s bill establishes a $0.15/mile vehicle miles traveled tax on commercial AV operations. Fifteen cents may not seem like much, but DC's gas tax of 35.3 cents per gallon pencils out to about 1.4 cents per mile for the average driver (assuming an average of 25mpg). Allen's bill would tax AV miles at roughly ten times as much as average drivers – far above the 1-3 cents per mile that other states piloting VMT fees (for human drivers) have proposed.

VMT fees are sound policy, and it's the right mechanism for an EV-dominant future where gas tax revenues are declining. A decade ago, I co-authored a report on state AV policy at the Eno Center for Transportation, which recommended AV VMT fees as part of a broader national strategy to address transportation funding challenges. But calibration really, really matters. An exorbitant fee risks discouraging shared fleet deployment while doing nothing to discourage the miles driven by privately-owned cars.

But it’s not just the VMT fees: The bill requires on-demand AV networks to comply with existing District for-hire vehicle laws. Today, Uber and Lyft pay a 6% gross revenue tax on all rides originating in DC, and Council raised that rate from 1% in 2018 to fund WMATA. As written, Allen's bill doesn't carve out an exemption from this existing tax, which means AV operators would likely owe the 6% TNC tax and the $0.15/mile VMT tax and multi-million-dollar permit fees.

Here’s some quick back-of-the-envelope math to illustrate the cost: assuming a 5-mile, $15 ride, an AV operator would owe roughly $0.90 from the TNC tax plus $0.75 from the VMT tax, which comes out to about $1.65 in per-ride DC fees – and that’s before we even amortize the $6 million in upfront application and permit costs. That's a significant cost stack that far exceeds what any human-driven rideshare pays today, and it may even prevent some AV companies from investing in the local economy and creating well-paying jobs for District residents.

DC Council should clarify whether the VMT tax is intended to replace or supplement the existing TNC tax for autonomous fleets. If it's both, DC risks discouraging shared fleet deployment while doing nothing to discourage the miles driven by privately-owned cars – which, I cannot emphasize enough, would potentially distort the market in exactly the wrong direction.

Revenue would flow into an Autonomous Vehicle Deployment Fund, split evenly between 50% for public transit infrastructure and services, and 50% for workforce development for displaced rideshare and taxi drivers. 

DDOT must also publish a workforce impact report within 180 days that analyzes the potential impacts of the deployment of AVs on taxi and rideshare drivers and proposes a spending plan for revenues collected in the workforce fund. But it’s unclear what new information the report’s analysis could uncover – and it’s also worth noting that taxi and rideshare drivers are still driving and making money in cities with mature AV services like San Francisco and Phoenix. Furthermore, under a 200-vehicle cap regime, DC's initial AV fleet would be a rounding error compared to the thousands of rideshare and taxi drivers operating on any given day.

If the Council is genuinely concerned about mass displacement, it’s important to remember that (a) the timeline for that concern is years away, and (b) the evidence from years of deployment in other cities suggests it may not materialize the way critics fear. Nevertheless, if DC plans to collect millions of dollars to address hypothetical worker displacement, it makes sense for the government to have a plan for spending it, even if the tax comes first and the plan comes second.

It’s also critical to note that introducing AVs into a city is not a zero-sum equation for workers: as Tiffany Moore from Waymo discussed at our April 8th event, operating an AV fleet requires physical infrastructure and human workers:

“There's an entire human infrastructure around our driverless vehicles — and those are local jobs for our workforce. When we think about investment, there are tens of millions of dollars of investment that would come with us having multiple depots in the District of Columbia to serve the citizens of the District.”

Half of the revenue flowing into the Autonomous Vehicle Deployment Fund is earmarked for public transit infrastructure and service. In practical terms, this means more money for WMATA – and as an almost-daily Metro rider, I welcome any new funding stream for the undisputed best metropolitan transit system in America (I will not be accepting any questions/comments on this). And linking some AV revenue to transit funding makes conceptual sense in dense metropolitan areas with strong public transit systems: if AVs are going to reshape urban mobility, some of the proceeds could help strengthen the backbone of the system they'll operate alongside. Reliable, high-quality, and high-capacity public transit helps to serve everyone better in urban environments – even those who do not take transit.

But the funding mechanism matters as much as the dollar amount. If the VMT rate is set so high that it discourages AV deployment or pushes operators to other cities, the transit fund will collect dust instead of dimes. A lower rate that actually attracts commercial AV operations to DC could generate more total revenue for WMATA than an aggressive rate that scares off investment entirely. Instead, DC should focus on collecting pennies for progress (to borrow from former Rep. Peter DeFazio). The Council should model this out, because the goal is to maximize sustained funding for transit and mitigate congestion while unlocking the many other benefits of AVs, not to maximize the per-mile earnings on paper.

The bill also takes a meaningful step toward multimodal integration, wherein permit holders operating on-demand networks must ensure their platforms are compatible with widely-adopted interoperable payment standards, and must collaborate with WMATA and regional transit agencies to explore a transit incentive program offering credits or discounts for AV rides that connect to public transit. I think it’s worth asking whether this framework could extend to Capital Bikeshare and other micromobility connections too.

This isn't just aspirational language: Waymo has already piloted exactly this model in San Francisco and Los Angeles, offering ride credits to passengers traveling to and from transit stations. If DC gets this right, AVs could function as a genuine first-and-last-mile complement to Metro rather than a competitor – which is the core promise of autonomous urbanism and arguably the single best counter to the fear that AVs will cannibalize transit ridership, while bridging the existing gaps in DC’s public transit network. 

The bill's liability framework was written with a trial lawyer’s precision: when an autonomous driving system is engaged, the AV company or ADS manufacturer – not the passenger, not a phantom "driver" – is considered the legal operator. The AV company / ADS manufacturer is responsible for traffic violations, tickets, and crash liability. Customers cannot be required to agree to arbitration regarding real or potential liability as a condition to use the service. ADS manufacturers must maintain a publicly accessible claims process.

This extends to consumer vehicles: any ADS manufacturer selling SAE Level 3 or higher systems in DC must register with DDOT, carry $5 million in insurance, and accept primary liability when their system is active. Few jurisdictions have attempted this approach for consumer vehicles – and it establishes a clear answer to the "who's responsible?" question that has plagued the national debates around personally-owned vehicles with driver-assistance systems.

Operators must file quarterly reports covering VMT breakdowns (by passenger-carrying, en-route-to-pickup, and repositioning miles), crash data, minimal risk condition events, and post-crash remediation steps. DDOT must publish these within five business days after receiving them. Crash notifications are required within eight hours.

One provision stands out, especially following last month’s Waymo/Waze announcement: AV fleets will be required to share digitized logs of infrastructure defects (e.g., potholes, damaged signals, pavement degradation), including GPS coordinates and timestamps, with DDOT. DC's roads getting continuous, sensor-grade inspection is a tangible public benefit that costs the city nothing to require – but actually collecting, storing, and analyzing the petabytes of data that AVs generate is no small feat and no small expense in terms of staff time and resources for both parties. 

Nevertheless, this could help DDOT identify local infrastructure issues and prioritize maintenance tasks – which could help to improve mobility for every road user and streamline public works projects.

Allen's bill is ambitious in scope and certainly has its flaws, but it reflects genuine engagement with the tensions around AV deployment from labor interests, trial lawyers, and other special interests – especially in a contentious election year for key positions in the District government. 

But there’s a lot of work to be done before this bill can deliver on its promise. The VMT rate absolutely needs recalibration: taxing shared AV fleets at ten times the rate of private cars doesn't discourage driving – it discourages shared rides in vehicles that (especially in Waymo’s case) are proving to be demonstrably safer than human drivers – and zero-emission, unlike the gas-powered cars they’d replace. 

The Council should also clarify how the VMT tax interacts with DC's existing 6% TNC tax because, as written, AV operators appear to be on the hook for both – which is a formidable fee stack that no other jurisdiction imposes. Plus, the $6 million in upfront application and permit fees that are due before a single ride can be offered will be a big factor in whether AV companies choose to invest in the infrastructure and workforce necessary to operate in DC – or skip it entirely. The Council should also take into consideration how the vehicle fleet caps will interact with its goals for equity and mitigating congestion.

The bill sets January 1, 2028 as the earliest definitive date a permit recipient could launch an on-demand ride network, but that would require the law to take effect by roughly spring 2027, giving DDOT enough time to stand up the commercial permit (180 days) and review an application (90 days). If the legislative process slips or the executive branch imposes further delays or requires further studies to be written (and DC's track record suggests all three are possible), then actual rides could push into mid-to-late 2028.

For now, the most impactful thing the Council could do during markup is recalibrate the VMT rate, clarify the fee structure, and assess how fleet caps and equitable service requirements interact. Get that right, and DC could create a national model for autonomous urbanism – but, get it wrong, and DC will have written the most comprehensive AV law in America… that no one ever operates under, preserving the status quo.

I'll be watching closely as the bill heads to a committee hearing. Stay tuned for upcoming pieces related to transit and multimodal integration, AV governance, and workforce considerations.