In the past post, I introduced the a factor of urgency into the concept of government by data based on the previous post that emphasizes frequent making of short-lived rules. I invoked urgency as a criteria in an attempt to place an upper bound on the number of rules active at any time. This attempt was to promote data by government as an answer to our current problem of having an excessively large number of laws and regulations. Having more frequent short-lived rules based on urgency can limit the number of rules in force at any time.
In government by data, the laws and regulations are short-term rules based on the latest data but there can still be a large number of rules involved. One of the reasons we tolerate a large number of laws and regulation in our current system is that the laws and regulations last a long time. Although burdensome for the governed citizen or corporation, there is time to learn all of the laws and make adaptations because the laws remain relatively stable for a long time. With government by data, the rules are short-lived and potentially radically different from one iteration to the next. Given this pace of rule making, there is an lower limit on the number of rules that the population can manage.
I proposed that government by data would limit the rules to the highest priority issues. I called this urgency. In the last post, I assumed that this urgency would come from the public. The public would demand certain actions to be included in the next cycle of rule making. As certain issues become more important, the older issues will become less important and they may be dropped from the list. I reasoned that the lower priority issues need no rules because the public has learned to tolerate these issues. I also supposed that short term rules can fundamentally change culture or attitudes to tolerate any lingering discomfort for some issue long after its rule expires.
Rules are a means to educate the public so that eventually they do not need rules to force them to do something. I made the example of educating the public about the importance of maintaining health-insurance coverage in order to avoid being denied healthcare insurance when they encounter a medical condition. A short term rule (for example, one that is effective for four years) may mandate everyone to have coverage. Once the population is educated of this practice, the rule may be removed and most people will continue to maintain coverage even though there is no rule mandating it. The result of the rule is to educate the public of the need to tolerate expense of continuous coverage even while they are healthy.
Government by data produces a very rapidly changing environment of rules. The challenge for the public is to stay current on the rules that they must follow. Only the most urgent rules will get attention. There is no value in having rules that no one pays any attention to. We could create low-priority rules, but they will be irrelevant if they are ignored. The actual set of rules in place at any time are the rules that have the highest urgency.
The problem is how to measure urgency. The best answer is that urgency itself is a rule based on data. Urgency is a rule that uses data to prioritize the possible rules into a set of rules that are most likely to be followed and that would most likely make a difference for present challenges. Urgency is a rule that governs the number of rules. Current data will identify major problems that need attention. Also data about the population can inform us of what combination of high priority rules will be successfully followed. In this definition, Government by data and urgency is redundant. It amounts to government by data but with an upper limit on the number for rules at any time. Some of that data informs the algorithms to effectively prioritize the rules to the optimal set to enact during the current iteration.
However, this concept of urgency is not what I had in mind as I wrote the previous posts. In those posts, I imagined a democratic approach to select the urgent issues. This may come in the form of petitions or protests. Or it came from elections to establish some mandate. I now think that democratic approach to prioritizing rules is not appropriate for government by data. Democracy even for the selection of most urgent rules will reduce the benefits of government by data. A point I made earlier is that data should obligate the population to accept and cooperate with the decisions. Allowing that same population to set priorities for rule-making will allow the population to avoid this obligation. In particular, setting the priority by majority vote will allow the majority to avoid obligations to follow the data.
Having a democratic approach to selecting rules to follow in a current iteration will result in a purer democracy than current governments that are democratic-republic government of democratically elected representatives. The data-driven rule-making eliminates any role for the representatives, and the urgency by majority selection of priorities will result in a pure democracy. Certainly, a pure democracy is an option for government, but is is different from what I’m suggesting as government by data.
I don’t find a pure democracy to be appealing because I agree with its characterization as rule by a mob. Mob-rule frequently involves passions that can cause the majority to oppress the minority. It is a possible choice for government, but it is not government by data.
To distinguish government by data from a democracy, the urgency of issues that require rules can not be based on democratic selection. The urgency itself must also be measured by data. Data can identify urgent issues as those that are presenting the highest risk of harm or the largest opportunity for benefit. Alternatively, data about the population’s sentiments and competencies can identify the most practical set of rules that collectively can accomplish a desired outcome in the current iteration.
Urgency provides a limiting factor on the number of rules. The population must not have a direct say in setting the urgency. Instead, government by data is a form of authoritarianism where the justification for the authority is the use of trusted data and the promise of rapid renewal of rules when new data becomes available. For any particular iteration of rules, the public must be obligated to follow the rules that are properly derived from recent and valid data.
In government by data, the public does not have the freedom to choose what rules to follow. That freedom to choose rules would result in a democratic or democratic-republic form of government.
I sought to define a new form of government that recognizes the represent-day challenges in the context of data science. The data science perspective recognizes the future as the source of observations that will inform decisions. In contrast, the standard scientific perspective recognizes causality where the present-day challenges are a consequence of the past. Government by data is government that is pays attention to the future instead of the past. The data are new observations emerging from the future. Government by data makes decisions solely on these observations. Government by data avoids past assumptions such as theories or other prejudices. Government by data requires nimble decision making that allows for no time for deliberative cognitive justifications for decisions. The rapidity and frequency of the data-driven decisions allow for responding to new data that may indicate an error or may indicate a new opportunity.
Government by data is for a new form of government that is distinct from older forms of government. The distinction is the central role of data in decision making. Government by data is definitely anti-democratic. Accepting a democratic role in decision making (even to set urgency) is to allow introduction of past theories or prejudices as data. Pure government by data rejects such forms of data because it accepts only confirmed direct observations of the real world ideally without any human prejudice or deliberation.
With the recent access to big data and big data technologies, government can evolve to make decisions solely on data instead of public debate. Such a government removes democratic participation from the decision-making process because data and analytics fully determine both the priorities and the specific rules. This form of government enforces the rules by policing the population to follow the current set of rules that do not come from democratic processes.
The rules are temporary based on the best data at the time the rule was made, and that new rules will replace current rules in a short time. The new rules will use the latest data from public participation in past rules. That data will include evidence of what was wrong with the older rules and this will lead to improvements in data sources or analytics. As the public learns about how government by data removes the possibility of capricious or arbitrary administration by fellow citizens, they will become more cooperative with the automated rule making.
There is a role for public democratic participation in this form government. This participation involves the public learning skills to query, to analyze, and to evaluate algorithms of the government’s data systems. The public does not participate in the creation of the actual rules. Instead the public participates in the underlying data-science project in order to improve it for future rules.
Public participation in government by data is through the use of open data and open source analytics. Although the decisions will come from the analytic algorithms based on the latest data, the public will have direct access to the same data and to the source for the algorithms. Instead of challenging the rules themselves, people can challenge the data and the algorithms. More specifically, the public is responsible for maintaining the processes that create new rules. Because all rules are short-lived (naturally becoming obsolete when new data arrives), the participation in improving the data and algorithms will provide immediate benefits by improving the near-future rule-making.
Also, with data analytics, the data effectively provides the algorithm for rule making. People can obtain more favorable rulings through organizing in ways that will generate new data that can steer the decision making algorithms. For example, the population can avoid unwanted rules through organizing to tolerate the issues that previously made the rules a higher priority. Similarly, the population can organize to emphasize conditions where they want new rules. In both cases, the organization produces new data that the algorithms will consider in making future rules that will replace older rules. The current set of rules are temporary and there is no obligation to refresh old rules if the data no longer supports them.
Although government by data gets democracy out of the making of new rules, this type of government can offer a stronger form of democratic participation in government. The public has direct access all data and algorithms used for making rules. The public can (and should) challenge any data or algorithms that are not trustworthy. Also, the public creates new data for future rules so that organized movements can change the data that will favorably influence future rules.
Government by data can be more democratic than the current system modeled on long-lasting party politics. Instead of expending energies in assigning power to one of the parties encompassing a broad coalition of issues, the population spends its time organizing around specific individual issues to obtain the rule-making they prefer and to avoid the rule making they dislike.
Active and aggressive participation to influence data can result in a very democratic means for self-government. Such power could ultimately have good or bad outcomes, just like any other kind of government. For example, it could result in organizing to influence rules that oppress some disadvantaged group. However, such an outcome would have to obtain a good score from trusted analytics of trusted data. The data and analytics must find some broad benefit from the next iteration of rules. If some dastardly oppressive intent is successful, it will need data to support a analytic finding of substantial benefit for the society as a whole.
In any case, this type of democratic control requires active and sustained participation to generate the data that will influence big-data statistics and in a way that is desired. Given the 3V data and the breadth of decision-making, there is a risk the modified data can result in unexpected rules in some other ways. For example, an attempt to influence the algorithms to choose a single-payer health insurance can result in an unexpected rule to move toward self-service health care (people provide care to themselves or close associates instead of using medical professionals).
See for example this simulation that demonstrates how analytics can transform a community based on some desired set of outcomes (scroll down to see the various scenarios). In particular, the ideal goals of the analytics can result in unexpected outcomes. This simulation is instructive also because successive examples shows modifications that can occur in successive iterations of government-by-data, leading to better algorithms for the future. Even within a single algorithm, the population can manipulate the outcomes by providing data for different preferences for diversity of neighbors. The user of this simulation is analogous to the population of a data-driven rule-making government in the sense that they have control over the data that will determine the future outcomes.
In summary, government by data is different from other forms of government because the identification and selection of enforced rules is completely automated by trusted analytic algorithms that rely solely on trusted data. This form of government can have democratic components through making the data and analytic source code open to the public. Also, the public can indirectly control the data through organization and public education efforts in ways that will influence future iterations of rule-making. Despite the automation of rule-making, this form of government may result in a new form of democracy that concentrates on specific issues to obtain desired future rules. This form of democracy may be more effective than the current approach that concentrates on what party has power to make rules through human deliberation.
12 thoughts on “Government by data and urgency, but urgency must be defined by data”
Pingback: Rethinking lessons from prohibition: illustration of government by data and urgency | kenneumeister
Pingback: Government by data, a different approach to separation of church and state | kenneumeister
Pingback: The role of charters in government by data and urgency | kenneumeister
Pingback: Spark data: distracting data deliberately introduced to influence analysis | kenneumeister
Pingback: Dedomenocratic Party in a democracy facing external enemies | kenneumeister
Pingback: Dedomenocracy illustration: safety of airline travel based on statistics not physics | kenneumeister
Pingback: Economy of compensated opinions in a dedomenocracy | kenneumeister
Pingback: Dedomenocracy: unsupervised government | kenneumeister
Pingback: Dark data: fall of civilizations | kenneumeister
Pingback: Rethinking lessons from prohibition: illustration of government by data and urgency | Hypothesis Discovery
Pingback: Government by data, a different approach to separation of church and state | Hypothesis Discovery
Pingback: The role of charters in government by data and urgency | Hypothesis Discovery