Improve Government With Algorithms - Without Politicians

12/17/2015 11:57 EST | Updated 12/17/2016 05:12 EST

There is no shortage of news to provoke apathy towards politicians if you're one to notice; in North America, the hypocrisies of political messianism are in full view. Barack Obama laments the intransigence of domestic gun control efforts while approving a $1-billion arms deal with execution-positive Saudi Arabia. Weapons are intended for their controversial military campaign in Yemen; accused of war crimes due to reckless bombing and high civilian casualties.

In Canada, Justin Trudeau is enjoying high approval ratings, embraced by progressives for gestures of inclusivity in selecting his cabinet and steps to reform marijuana laws. A docile Canadian public mostly ignoring his strange, reactionary plan to exclude single heterosexual Syrian men from the Liberal refugee scheme; and his decision to back former Toronto Police Chief Bill Blair as a star Liberal candidate in the election -- the man responsible for "the most massive compromise of civil liberties in Canadian history," according to Ontario Ombudsman Andre Marin.

Algorithmic governance is a radical, digital reimagining of government centred on computerized processes.

The most abused cliche in politics is the concept of 'change,' yet a young movement among academics and techno-scientists seeks to overhaul the current system with a computerized, politician-minimal alternative. Algorithmic governance is a radical, digital reimagining of government centred on computerized processes. Algorithms -- which already have many applications like sorting incoming emails and controlling traffic lights -- would be unified to create a governing network.

Enabled by big data, algorithmic governance is a pure expression of technocracy and celebrated as a path to enhanced society: "we are at the beginning of a big data algorithmic revolution that will touch all elements of our society," writes digital media baron Tim O'Reilly, in Beyond Transparency, a book of open data musings. In contrast to the current system -- accused of being byzantine, asphyxiated by rigid and outdated regulations -- advocates like O'Reilly describe a more efficient and adaptive model of government that would be less political, comparatively free of human inefficiency and possess real-time flexibility called "ultrastability."

Algorithmic government may sound far-fetched, but it is already happening in smaller, more localized ways. The U.S. Securities and Exchange Commission (SEC) uses algorithmic devices to detect suspiciously high performing hedge-funds for potential trading violations. In Italy the National Revenue Agency has implemented a tool called "redditometro" that uses algorithms to analyze personal spending and look for discrepancies that suggest tax fraud. Ford Motor Company and Intel are developing facial recognition software that will require car's system to authenticate all operators before it can drive. On a bigger scale, Singapore is the first country to experiment with macro applications of algorithmic governance.


Otto Gustav Carlsund (1933)

Techno-scientific solutions to social problems is not a new concept, but appetizing in today's dismal political climate of bigotry theatrics and machismo nationalism. There is a modest admission to the limits of human capability in all of this; a kind of sensible, informed misanthropy. Yes we can't. If computerized government suggests an antidote to cult of personality and strongman politics, advocates are correct to acknowledge the limited human ability to govern and regulate itself; a message that echoes the Art Concret movement of the 1930s, which rejected the deity of the artist, embraced anti-expressionism and scrubbed work of the problematic human tendency towards symbolism and sentimentality. Coping via erasure. Artists such as Theo van Doesburg and Otto Gustaf Carlsund used geometric processes and cumbersome materials to make non-referential work that minimized the artist's touch and clarified the work. Both movements share an idealism wrapped in measured self criticism; radical honesty for ambitious objectives.

Optimism aside, advocates should expect resistance and claims of intrusiveness.

The discussion around algorithmic governance is young but it's exciting, even romantic to picture election campaigns as contests between code bundles; identity politics and sleaze minimized, absent rich, old men in makeup as aspirational paragons of achievement and the comedy of partisan discourse. It's already possible to envision O'Reilly's moderate vision of a digitized regulatory sector versus a complete rewrite of government.

Optimism aside, advocates should expect resistance and claims of intrusiveness. For example, the use of GPS satellites to automatically issue driving offences without a police officer ever pulling over the driver could feel Orwellian. Healthy skepticism is justified after some early mishaps. The "Flash Crash" of 2010, a trillion dollar stock market crash that lasted less than an hour, crippled the S&P, Dow Jones, and Nasdaq 100 and was the second largest intraday stock market point swing ever. Subsequent investigations determined it was triggered by a perfect storm of big trades, financial "spoofing algorithms" and a predominantly digital, self-activating system; a demonstration of the 'automatic' and emotionless nature of algorithms that critics assert. Also, determining accountability after an algorithm goes wrong is unclear: who ultimately is responsible for a codified process?

Scholar and tech-skeptic Evgeny Morozov cautions that proponents of algorithmic governance aren't envisioning a smaller or less invasive state but a more controlling one. Their distillation of complex problems to simple measures, sensors and feedback loops -- which he calls "solutionism" -- is naive and a "concept at odds with the vocabulary of democracy." Mission creep is also a risk if collected data is used for malicious purposes like the NSA PRISM surveillance program, so information safeguards would be crucial.

Critical voices should be embraced, but as computer scientist Adam Elkus writes in Slate, many of the problems projected onto algorithm governance already exist. Criticisms reflect "an age-old argument about impersonal, automatic corporate and government bureaucracy." Supporters insist a digitized system would alleviate democratic abuses and provide non-hierarchical government that is less susceptible to strongmen and free of bureaucratic entanglements. An improved model of governance braided with a healthy streak of the Concretists' self-censure.


Trudeau Greets Syrian Refugees