Recent Thunks

New York, August 20, 2015 - Churchill nailed it in 1947 when he said “democracy is the worst form of government, except for all the others which have been tried from time to time.” For decades thereafter traditional Western Democracy was the tarnished global standard.

But with the Internet’s ascent people began to believe something better was possible. That belief fueled our advocacy for the .nyc TLD. Through it we imagined more open, transparent, and accountable governance processes, a post-Churchillian democracy where connected residents could readily participate in decision making. One foundation for that was a civic commons.

[Background: The term commons refers to cultural and natural resources accessible to everyone, such as air, language, water, and a habitable earth. Commons are not owned by the private sector or government. With the activation of the .nyc TLD a new civic commons became possible, one formed around domain names such as issues.nyc and voting.nyc. More on this below.]

A lack of meaningful support for .nyc from City Hall diminished our expectations for our city’s TLD and for the civic commons. In recent months, we’ve been turning our attention elsewhere. But two recent reports highlighted the need for a civic commons and jerked our attention back to .nyc. The first was an article in Science reporting on research demonstrating how the Internet exerts a massive influence on elections. The second, a New York Times article detailing the opacity of algorithms (invisible computer code that determines who sees what in searches), and how they result in the de facto placement of gender and racially biased ads.

Taking a fresh look at these and other challenges brought on by digitization we decided to present the role and advantages a robust digital commons would have on our city’s operation and quality of life. But first a look at those reports in Science and the Times.

Biased Algorithms

The Science article reported on work by Robert Epstein that conclusively demonstrated what most of us know intuitively: The higher a candidate’s rank on a page of Internet search results, the more likely voters are to choose them. Summarized his finding Epstein said: “What we’re talking about here is a means of mind control on a massive scale that there is no precedent for in human history.”

That’s a strong claim. Let’s take a look at his research.

In one experiment, Epstein, a research psychologist at the American Institute for Behavioral Research in Vista, California, and his associates recruited three groups of 102 volunteers. The researchers built a fake search engine they called Kadoodle that responded to search inquiries by returning a list of 30 websites, 15 for each of the candidates. What the volunteers didn’t know was that the search engine had been rigged to display biased results. For example, in one scenario a subject would see a list of 15 websites with information about one candidate followed by 15 similar results for the opponent. Predictably, the subjects spent far more time reading Web pages near the top of the list. Before and after questionnaires detailed the impact of the biased presentation: in one instance the rigged search results increased the number of undecided voters choosing the favored candidate by 48%.

Interestingly, the subjects who noticed they were being manipulated were more likely to vote in line with the biased results. Of these voters Epstein says, “What they’re saying is, ‘Well yes, I see the bias and that’s telling me … the search engine is doing its job.’”

In a second experiment the scientists recruited 2,100 participants through Amazon’s Mechanical Turk. The large sample allowed them to pinpoint the demographics of those most vulnerable to search engine manipulation: the divorced, Republicans, and subjects who reported low familiarity with the candidates. From these results, Epstein concluded:

“In a two-person race, a candidate can only count on getting half of the uncommitted votes, which is worthless. With the help of biased search rankings, a candidate might be able to get 90% of the uncommitted votes [in select demographics].”

Other research by Epstein revealed that a search engine doesn’t have to intentionally manipulate for a biased effect to manifest. Search algorithms put one candidate’s name higher on the list based on factors like “relevance” and “credibility.” But the meaning and impact of these terms on displayed results are closely guarded by the developers at Google and other major search engines.

“It’s easy to point the finger at the algorithm because it’s this supposedly inert thing, but there are a lot of people behind the algorithm,” commented Nicholas Diakopoulos, an independent computer scientist at the University of Maryland, College Park. “I think that it does pose a threat to the legitimacy of the democracy that we have. We desperately need to have a public conversation about the role of these systems in the democratic processes.”

The New York Times article, When Algorithms Discriminate, talked about two forms of biased search algorithms that deter advancement by women into executive positions. In one instance job search sites offer more good jobs to men than to women. In another image searches for CEO’s exaggerate men’s occupancy in top positions.

Upon close examination it’s clear that some code is negatively influencing our fundamental rights and beliefs. In one instance decades of progress on expanding and enhancing the quality of the vote is being lost by the replacement of traditional candidate evaluation systems by new technology. In another, gender and racial equality are silently being eroded by invisible filters that insufficiently respect our cultural norms.

We need to be more vigilant about protecting our hard won standards for human and social rights. Luckily there’s a commons within the .nyc that can assist our addressing these and other civic issues

The Civic Commons

  One of the empowering features of a city TLD is the new “space” it creates. It’s not as tangible as the land upon which housing and office space are created, but in a digital era it offers enormous opportunity. The commercial world was quick to take advantage of digital spaces. It’s time civil society identify and occupy commons space within the .nyc TLD. Here are a few examples of domain names that will support a civic commons:

• Search.nyc - To address the voting issues discussed above we should craft the search.nyc domain name to provide fair returns on candidate searches. Designed with impartiality and transparency in mind, it can help foster fair and trusted elections.
• Voter.nyc - The voter.nyc name set (voter.nyc, voting.nyc, voters.nyc, candidates.nyc) is another part of the civic commons that should be crafted to facilitate elections and election time decisions.
• Issues.nyc - Long and short term discussions of citywide import should be addressed in an issues.nyc space. These could feed into the voter.nyc name set.
• Neighborhoods.nyc - A most useful set of commons spaces is being made available through the neighborhood names. Nearly 400 names have been set aside to enable local residents to address the concerns of everyday life. Never before have New Yorkers had access to an effective local communication system.
• Meta Names - Intuitive names that facilitate locating these sites need to be identified and developed, for example, CivicCommons.nyc, NewYorkCommons.nyc, and CommonsIndex.nyc.

In creating our civic commons we can build upon the success of those in other realms such as fisheries, farm irrigation systems, and Wikipedia. Nobel Prize winning economist Elinore Ostrom detailed four factors that define areas where commons can succeed, with a city TLD meeting them all.

• A definable boundary. This is certainly true of intuitive domain names. Each is a unique name within the well defined boundaries of the domain name system. Their intuitive nature, coupled with effective marketing, makes
• A second closely related factor is difficulty of substitution. Once can easily create a bestsearch.nyc or importantissues.nyc, but once claimed and coherently marketed as the city’s patrimony, it becomes clear that there is but one commons.
• The third is the presence of a community. While a big, great city, New York is but 1/10th of 1% of the world’s population living on 450 square miles of land. Its residents have a common interest in assuring the social and economic betterment of their city.
• A final condition is that there be appropriate community-based rules and procedures in place with built-in incentives for responsible use and punishments for misuse. Crafting these rules is the primary challenge; but with precedents from the likes of Wikipedia, the task should be doable.

Creating The Civic Commons

The planning, design, funding, and oversight of the civic commons will best be achieved using a multi-stakeholder engagement model involving academia, business, civic society, government, residents, and the technical community.

There’s good news here. The de Blasio Administration took some positive steps early on by setting aside some commons names on a reserved list. A preponderance of reserved name were neighborhood names - GreenwichVillage.nyc, Harlem.nyc, etc., but search.nyc was on that list as well. City officials are currently looking for a business and governance model to support the operation of the neighborhood names. The adventurous can go today to neighborhoods.nyc and begin the process of acquiring a neighborhood domain name.

Making the commons a success requires fine tuning a governance model. Wikipedia and other successful commons provide some lessons here. But many refinements will be required before they succeed. There are also design and technical issues, especially with search.nyc.

But success with the commons requires support from City Hall.

• City Hall must show leadership and promote them with the same vigor and persistence used for 911, 311, and nyc.gov. As an initial sign of support it should move the city’s website to gov.nyc.
• It needs to invite the stakeholders into the planning processes.
• And it needs to relinquish its grip on their operation. The stakeholder communities that oversee the various names must have ultimate say in their operation.

With .london, .paris, .tokyo and 30 other cities needing similar search tools, the mayor should strive for a collaborative development process. Mayor de Blasio should  announce that New York City will be establishing a global standard for a city search engine and the commons. The city should issue a Request For Proposals to identify a responsible operator of this important resource. An independent funding stream is needed and a governance process enabling all the stakeholders to participate in its development and oversight.

Read more:

• Robert Epstein’s experiments in Science are available here.
• See this Times article on gender and racially biased algorithms.
• See our Transparent Search wiki page for starting points on building a city-friendly search engine.

—-

Our thanks to KraljAleksandar for Churchill, Jagarnot for Syndicated Mind Control, Victor Ponce for Tragedy, and Christine Prefontaine for What.

all images made available through Creative Commons licenses.