Last Sunday, The Washington Post published my op-ed criticizing the approach taken by Councilmember Mary Cheh’s introduced legislation to legalize driverless cars in Washington, D.C. Briefly, I argued that Cheh’s bill unduly restricted testing and the wider future adoption of autonomous vehicles via three requirements: (1) a licensed driver must be in the driver’s seat, (2) driverless cars must be “alternative fuel vehicles,” and (3) autonomous vehicles would be subject to a new vehicle-miles traveled (VMT) tax.
I noted that this all betrayed a certain dangerous attitude to technological innovation, albeit one that was well intended. “But no one knows precisely how autonomous vehicle technology will develop or be adopted by consumers,” I wrote. “Cheh’s bill presumes to predict and understand these future complexities and then imposes a regulatory straitjacket based on those assumptions.”
As Adam Thierer, a senior research fellow at the Mercatus Center at George Mason University, remarked, “The [driver's seat provision] is the one that most closely resembles the traditional Precautionary Principle, but the other provisions are based on a similar instinct that progress can be preemptively planned.”
Cheh has responded to my op-ed with a letter in the Post. In it, she says she has removed the alternative fuel and VMT tax provisions. Although the District’s Legislative Information Management System still shows her original bill in all its terribleness and we lack access to Cheh’s private thoughts and documents, I’ll take Cheh at her word that she plans to amend her bill in the right direction.
However, Cheh’s letter reveals a more serious, underlying problem. “Perhaps Mr. Scribner is unfamiliar with the legislative process: After a bill is introduced, legislatures hold hearings and invite comment in order to improve the legislation before it comes up for a vote,” she wrote.
Au contraire, Councilmember Cheh. I am indeed familiar with the legislative process, and not just in D.C., but in other jurisdictions as well. Prior to Cheh’s bill, three states (Nevada, Florida, and California) had passed legislation that legalized autonomous vehicles and they all had something in common: none included bizarre alternative fuel and VMT tax provisions at any stage of the legislative process. All of these bills had passed legislatures well before Cheh introduced her bill in D.C., so one would think a prudent legislator would have familiarized herself with current legislative developments instead of introducing bills that invite derision.
No such luck. Cheh waited for a hearing to be told these obviously problematic provisions might in fact be problematic. Noticeably absent from those testifying was Google, which has developed the most advanced autonomous vehicle prototype (that we know of), the Google Self-Driving Car. Two automakers experimenting with autonomous technology, Volvo and Daimler, were present.
Cheh, however, doesn’t seem to understand that there are broadly two competing autonomous vehicle technology models. The first is what I would call “true autonomous,” such as Google’s prototype, which does not rely on roadway vehicle-to-vehicle communication to sustain operation. The second is connected vehicle control, which Volvo and Daimler are backing, that relies on inter-vehicle communication for operation.
Cheh defends her California-style prohibition of truly driverless cars by quoting a Daimler lobbyist:
In a few years, cars will be able to operate reliably without drivers, but we are not there yet. As the representative from Daimler testified at the hearing, “We do not believe we’re at the point of replacing the driver. . . . And I suggest we not rush into it, because the public streets really aren’t the place for a ‘trial and error’ testing approach.”
First, no firm is likely to offer any service without a driver unless it safe to do so. There certainly won’t be any consumer models on the roads until safety is assured. The driver’s seat provision that was adopted by California and now championed by Councilmember Cheh was meant more to pacify the irrational, risk-illiterate fears of politicians than to actually protect the public. Cheh has not even proposed creating a pilot program to test true autonomous vehicles sans licensed driver.
Second, observers of driverless car development, including myself, have expressed concern that connected vehicle technology is too infrastructure intensive, meaning that widespread adoption will be dependent on costly government roadway upgrades. Indeed, one of the main technical benefits of true autonomous vehicles, such as the Google Self-Driving Car, is that they can operate on existing roads in existing traffic conditions. Another advantage is that independently controlled, true autonomous vehicles face lower manipulation risk (this is the digital age, after all) presented by either government or criminal hackers.
Depending on the changes to Cheh’s bill that actually materialize, I, as a D.C. resident, could be inclined to reluctantly support it even with a limited temporary requirement that a licensed driver be in the driver’s seat. But Cheh’s present bill should serve as a warning to other legislators: don’t presume to know how autonomous vehicle technology will develop and be adopted, and don’t write bills based around these speculative and probably incorrect assumptions.