Artificial intelligence can be curiously stupid. My Android phone still thinks I’m “wing Cruz” and doesn’t know my kids. Pandora overplays The Church and Deadmau5 (no offense).
As hackable as Alexa, Jeep Cherokees and credit card services are over a public Internet not designed for security, the networked gadgetries enabled by the Internet of Things (IoT) continue to dazzle. An overwhelming number of them will be on display again in January at the 2020 Consumer Electronics Show (#CES2020).
The primary vulnerability of the IoT is not hackers, though, but IoT policy.
Technology that overcomes “market failure” in the provision of goods and services should enable the reduction and streamlining of regulatory burdens. Instead it sometimes threatens to foster the expansion of government power.
That is, the same IoT that animates objects can also mean instantaneous nanny-state regulation from a distance—of drones, vehicles, buildings, social media use, schooling and more.
You’ve heard of free-range kids vs. helicopter parenting?
Well, the society fancying itself on the verge of flying cars may face helicopter government instead; assorted bureaucrats clicking and swiping from afar, using the IoT to control the IoT.
It’s one thing for Tesla to send its own software updates to its customers’ cars. We definitely want such things to happen—a lot.
But as Jason Dorrier noted in Singularity Hub, "regulations … written into software” could be highly appealing to regulators. A “No drones within 100 feet of federal buildings,” rule, for example, could be enforced by requiring the uploading to networked objects of software patches altering GPS coordinates, and disabling them in event of non-compliance.
Dorrier named other examples: software patches imposing speed restrictions and no-drive zones on vehicles, preventing cars from starting without seatbelt attachment, and mandating thermostat settings and water use restrictions in buildings.
Entrepreneur Marc Andreessen long ago described software eating the world. Titans in sectors from “movies to agriculture to national defense” are now software companies, run on software and delivered as online services.
Unfortunately, while software has eaten business models, it is not eating traditional top-down central regulatory regimes in the sense of displacing them.
Those systems are preparing to eat the IoT instead.
The next step in this “evolution” could go beyond rules mandating the updating or patching of software, to unelected bureaucrats simply doing it themselves remotely by clicking and swiping rather than enacting a law or rule, Constitution notwithstanding. The use of guidance documents, informal directives and other “offers you can’t refuse” are already a prominent regulatory concern highlighted by the Administrative Conference of the United States. The IoT could magnify such abuse.
In Long Beach, California, water authorities used smart meters to nab alleged water-wasting villains. The “logical” next step is government remotely overriding commercial building water and energy usage, perhaps even drafting the utility to countermand people’s own smart home preferences. This is particularly galling since it ignores governments’ culpability in the prevalence of shortages.
For policymakers infected with Green New Deal (GND) fever and taking steps like installing rooftop-mounted carbon-footprint monitors, one need not speculate on how authoritarian they are prepared to be when it comes to tweaking office buildings and regional tailpipes. Google’s Street View cars fitted with chemical sensors now map air quality in several cities and provide that information to governments. The Environmental Protection Agency (EPA) has already invested in technology tracking office workers’ energy and water usage, complete with visual alerts.
The EPA has protested exemptions to the Digital Millennium Copyright Act that might permit owners and researchers to access and modify vehicle software on the grounds that emissions might increase. But one suspects a future GND-lubricated EPA would itself love to remotely modify emission control software in vehicles – but punish you harshly if you do the same.
Politicians like Sen. Josh Hawley (R-Missouri) take addiction psychology seriously, and Hawley has introduced legislation to require Big Tech to eliminate “infinite scrolling” and set timers on gaming and social media. Future rules would be written and enforced by the Federal Trade Commission under this (so far, unpopular) scheme. Since these services are cloud based too, helicopter government may elect to ensure future compliance should such a regime emerge. For the children, of course.
In a progressive society bent on turning professors and administrators into government employees via “free” college, major textbook publishers are going digital. So do not be surprised if students are required to use the government’s preferred “constantly updated texts, tethering students and schools exclusively to the publisher’s digital platform.”
And so on.
The “Smart Cities” movement potentially may present the most significant incarnation of Regulation from A Distance, though. Versions I’ve noticed seem to involve heavy governmental control, "partnered" with large corporations or government-preferred players.
Despite inherent operational difficulties like procurement challenges, public-private partnerships indeed can play important roles in society (the more open to transition to fully market operations they are, the better, in my view). But “PPPs” present problems if they trend toward the “public” piece running the show and locking in a permanent regulated public infrastructure and utility mindset rather than ever envisioning any path to voluntarism and privatization. Here especially, the very IoT technology capable of overcoming ancient market failure justifications for government regulations and steering can be misused to expand the administrative state instead. As in any field, the political entrepreneurs among Big Tech’s ranks will be happy to oblige and offer the licensed/approved technologies, such as traffic routing, vehicle to infrastructure V2I communications, securing a monopoly on emergent local or regional drone deliveries, or scoring coveted e-scooter contract (while I can’t park my motorcycle on the sidewalk).
It may be perfectly reasonable for Uber and other ride-sharing outfits to collect and share vast amounts of anonymized locational and traffic data with public authorities. But for reasons apart from the emissions tweaking noted, authorities in “smart” cities are also attracted to remote override capabilities over automated vehicles for clearing lanes for emergency vehicles, routing around police and fire activity, or shutting down the vehicle of a fleeing suspect. On an already insecure-by-design Internet on which targeted hacking could cause monumental stranding and gridlock (or even weaponize vehicles) this global “LoJack” is concerning and feeds into a suspicion that smart cities will be primarily surveillance cities that ignore requirements for court orders, echoing abuses of red light traffic cameras and biometrics. In China, many vehicle manufacturers have been sending locational information to the government without the vehicle owner’s awareness, and facial recognition there enables jaywalkers to be outed on giant public screens.
Ironically, local city planners themselves sometimes game the useful traffic routing apps commuters rely upon to thwart them, and even create gridlock to drive distressed people into biking or public transport. Smart city planners can be similarly suspect.
The most significant issues in the IoT, properly, entail hazards or injury to third parties. But government can aggravate those. The federal government pursued a V2V (vehicle to vehicle) mandate that would have locked in obsolete technology and increased an array of risks across networked infrastructure. Furthermore, governments are prone to inappropriately indemnifying favored industries from the harms they cause, such as homeland security legislation that indemnified manufacturers of security technologies (like weapon alarms and bomb detectors) for losses above insured levels in the event of failure. Similarly, in the nascent space commercialization sector, the Spurring Private Aerospace Competitiveness and Entrepreneurship (SPACE) Act of 2015 (in Sec. 103’s extension of “Indemnification for Space Flight Participants”) absolves launch providers from catastrophic losses or for injury to third parties (through September 2025). In this vein, it is likewise interesting that the orbital space debris that has become a concern has happened primarily without a large private sector presence in space.
So in spite of the default to government that is now apparent in IoT, continuous safety and due care advances are not a challenge for industry, they are the normal goal and operational mode. Big Tech has lives and assets to protect.
Indeed, technology helps expose regulatory malpractice and expands the case for separation of tech and state. But it will be difficult to unspool opportunistic government entrenchment like that sought in the IoT if the bureaucratic impulse is allowed to prevail. Instead of worrying about autonomous vehicles, we’d best thwart autonomous regulation and ensure ample opportunities for opt-out, evasive entrepreneurship, and alternative business models.
And for goodness sake stop trying to do everything in the city. In colonial America, settlers sailed up the James and other rivers and built cities once they hit the rocks. Now the cities are the rocks. Rather than jump the gun with VTOL (vertical take-off and landing) air taxis, for example, practice in the county and the peripheries on networks of private land and construct altogether new property rights/corridor/route and infrastructure regimes. Then, sit back and enjoy the emergence of “new” cities out there. Afterward take the lessons learned and improve the old-school cities. Above all, do not use the IoT to “solve” a “city problem” and inflict it on the nation. Offer it as one option, sure, but not the only.
Foundational issues like these need to be worked out experimentally, not imposed push-button style. Scientists debate whether the best approach to AI is to program everything in, or to let it learn like a baby, or some mix. Routing by AI algorithm can save lives in airspace and highways by eliminating the human-error hazard, but it doesn’t have to be and ought not be government dominating this. I’ve joked that technology and tracking could eventually make it possible to pack the sky like a neutron star with commercial and personal drones, with defined corridors respecting rights. Data sharing by entities like the Space Data Association’s satellite operators plays a large already prevalent role in risk management and situational awareness in space with insights that could potentially cross fertilize what happens on the ground with cars, and in-between with drones. Instead, drones are being absorbed into decades old commercial airspace models, not to mention already required to be registered with the federal government.
Contractual devices, insurance and liability innovations that mitigate risk become easier, not harder in the normal course of events. Will those smart cites consist of hive-mind “roadway management systems” and networked cars rather than actually independent autonomous vehicles that detect on their, not imposed, terms? Not so far. Instead, fusion of the (regulated) vehicle and the (hyper-regulated) infrastructure is the path we are on now, and it may or may not be the right one, depending on the setting (cities vs. country, for example). Amazon has patented a networking technology to allow self-driving vehicles en bloc to navigate reversible lanes. Amazon, not the automakers, would own and operate such a network and presumably let any makers vehicles participate, and it seems a better, intermediate, approach than governmental management.
But bottom-up and top-down approaches need to remain viable, because what governments seem to want is not localized, independent control from inside smart, autonomous vehicles (and intermediate-scale networking innovations atop that), but external control over everything. It’s noteworthy that plain old ice-skating rinks thrive without somebody with a megaphone telling everybody where to go, and the same can be true over tomorrow’s networked world if allowed. Maybe it’s best for city planners to control roadway signals, rather than potentially individual vehicles or fleets of them.
Rather than top-down push button rules for the most complicated cases of cities, the rocks-in-the-river experimental stance also preserves the ability to explore trespass, peeping tom and vicious animal-style rules against misuse of drones and other IoT-enabled vehicles. as entrepreneur John Chisholm described in defense of “organic” rather than imposed regulation in Unleash Your Inner Company. Heck, given the pace of miniaturization, maybe there won’t be smart cars at all, but husks or “sleds” that we snap our mobile devices into. But that can’t be the case if smart cities amount to the major automakers teaming up with planners for a lock on the market.
By this point it should be clear that another major problem with remote regulation at a distance is that it steamrolls over synergies with sibling and cousin network infrastructures that should be evolving in parallel. The planners’ approaches so far preserve the siloed regulation of infrastructure that prevails today (power, water, sewer, transportation). The expansion of 5G and air and land corridors invites huge possibilities for simultaneous infrastructure development like privately managed drone docking/recharge stations (such as on lampposts and church property) and other heavy-duty networking assets. But such symbiosis isn’t happening that I can see.
The gaps and sub-optimality imposed by the threat of top-down, siloed IoT regulation are symptomatic of the myth that federal agencies offer expertise, let alone the myth of impartiality. Expertise in the modern IoT context necessarily consists of extending institutions of complex property rights and contract on land and in air and space. Bureaucrats, as well as the entire smart city concept, ignore that pretty much entirely in favor of the default public utility model as the backdrop and framework.
The concerns preoccupying IoT policy seem not to be the actual complex property rights problems needing resolution, but rather seem self-reinforcing of regulatory bureaucracy, even cronyism. The Internet of Things also needs to be the Internet of Thinks; that means not entrenching dumb, permanent market-socialist policy via misuse of remote regulation and pushbutton bureaucracy.
A main point of Jason Dorrier was that, while software updates typically add functionality, they can also take functionality away. While I might fantasize about pushing a button to ground Leonardo Decaprio’s next high-carbon-footprint jet flight to a global warming confab, I’d better not.
But the same goes for bureaucrats when it comes to All Things IoT; they’d best not. We’ve a long way to go to get even the easy stuff right; take it from wing Cruz.
Originally published at Forbes.