More on Public Services
Next: Public Side Architectures
Thematic Organisation of Services
After a longer pause, again found some time for substacking.
In past two posts we’ve discussed examples of public services - part 1 and part 2
Let’s return to the discussion on how and who creates the services.
When services are digital, I do not know who provides them nor how they are implemented as a user. I just shout at the cloud and a number of actions happen.
This means that there could well be a number of providers behind the service interface. For example, in health different diagnosis services could be implemented by different providers. These could be a public health organization, a private corporation, a patient association or even a self-owning fund based on open-source implementation. And a single service can have multiple implementations from different providers. Users can be given the right to select the service provider they like from these. Or a machine learning model can consult several ones and create a combined recommendation based on them.
Why would this be beneficial?
Let’s assume I have diabetes and there is a global company simply focusing in diabetes. They have customers from multiple parts of the globe with small differences in DNA, body structure, cultural eating and general living habits. Being tightly focused and having wide reach allows this enterprise to constantly improve the treatment for just this one disease.
Additionally, when more and more people from different backgrounds move around the globe, this specialised provider is in a good position to have a well-suited, personalised treatment whatever the patients’ background are. Something a national provider may lack or have more limited exposure to the multitude of differences.
This gives advantage to the specialist but does not mean that they would naturally be selected out of essentially a political process.
Some people like to see public organisations being responsible for organising and running the services. They fear that since public services today are natural monopolies that a single or few large private organisations can first with aggressive pricing and acquisitions reduce or eliminate competition and later start exploiting people who do not have a real choice of selecting providers as alternatives no longer are there or prices are set inside a small oligopoly. Public organisations do not have such monetary motivations.
Private provider models works worst if public side hands over the implementation responsibility to a single private organisation after a price competition. After winning, the objective for the provider becomes to provide services with the least amount of cost and since the winner had the lowest price, the worst possible service level using literal reading of contract terms is guaranteed. Adding insult to injury they may look for additional income by monetising the patients’ data in some way.
If on the other hand citizens are free to pick provider (given cost constraints) and public side also providers some services (to understand general price level), the risk is smaller.
There is naturally a valid care for private providers. It is based on the fact that public services are in monopolistic position and have no competition to drive quality improvements or to reduce costs by automating services. So, in places where management is of low quality, the public services do not evolve with time and can sometimes degrade significantly. To add insult to injury, people for managerial positions may be picked more based on party loyalty than experience on the field or ability to lead complex organisations. All that is needed, is to set the criteria for top positions emphasising proven track record to engage and influence domain specific societal dialogue, knowledge in relevant legislation + whatever similar comes to mind and the road is paved to appoint politicians who have lost public trust (and their seat) to lead key organisations and start filling other positions with friends from same party.
These are not the only alternatives. There are also constructs such low profitability limited liability corporation (L3C; or in UK Community Interest Company). Companies that have stated beneficial goal and can only generate limited profits. This eliminates the motivation to exploit people in vulnerable positions even when the company is private. Patient organisations and associations could well start L3C corporations to serve their members.
There is no natural law that prevents competition among public organisations. One simply needs to split public organisations into several competing parts and let them all serve customers with the same service mix. Competition drives them to work efficiently and automate what can sensibly be done by computers. Or on municipal services, allow any municipality to compete with any other municipality on digital service provision.
So, there is no technical reason for mixing and matching public and private digital services as long as they all implement the same interfaces and integrate to the same common infrastructure. On the cloud everything is behind an API call anyway and this masks the provider behind eliminating the need for monopoly providers.
Even public services between nations could be interchangeable. As a Finnish citizen I might opt for the Estonian eResidency service as my global identity provider and use it when I interact with Finnish state or municipal services. Or I might opt for an approved blockchain based identity provider.
There may well be some services that by nature are so central to how the state operates that speculations above do not apply. Services like:
Tax collection
Identity (national state issues identity documents that are the root of trust for other identity providers)
Ownership records (who owns this piece of land, shares in company or apartment)
Personal data records (even if health is provided by a 3rd party, the data itself might be best to be stored inside a sandbox provided centrally).
Statistics service (does not rule out 3rd party statistics via open APIs)
Law making and enforcement
Safety and operational reliability of society
But some of these do not necessarily need to come only from my government. As mentioned I could use the eResidency service provided by the Estonian government, the analytics could come from another EU national service provider even when the data would be stored to Finnish personal data storage service.
Today nations solve problems with organisations that are geography-based. The above organises problem solving based on the need or theme (health, education etc.) where approved solution providers do not necessarily need to be tied to national borders. You could call this thematic organisation of services.
Lastly a few thoughts on service creation.
Program Your Own Public Services
Today there is a strict division between public and private services and both sides make their own services and service flows. The users’ role is only to consume them.
Why not decentralise service creation and allow anyone to tie together the services from both public and private sphere into a flow that solves their problems.
In the IT sector there are three main ways to decentralise functionality or ‘policy’ to users.
First one is to have pre-created “recipes” or workflows where the user can set personal parameters. Let’s say I have solar panels; I might want the system to read calendar of all family members and automatically sell energy when we are on holiday and put the money to voluntary pension fund for kids. The parameters would be the access details for the calendars.
Second more flexible is to use tools where in a graphical programming environment one can “program” from Lego-like boxes the flow of a service. They’ve never become too popular as they work well for relatively simple cases and programmers tend to be text oriented. However, they do provide a level of flexibility that opens up personalization to people who do not have years of programming experience.
Example what a flow diagram could look like.
For what a real implementation would look like see: https://www.pipes.digital/docs
Third most flexible is to open the APIs and allow anyone to use any programming language like Python to tailor own services and share them via open-source code repositories.
If the interfaces to public services are easy to use and there are platforms that make tying them together easy enough, it becomes possible to anyone to tailor their own mixed services. Program their interactions with public and private players to meet just their personal needs and share the results to others who might have same or similar needs. All this without asking any permissions or telling anyone in advance.
In future large-language-models that can generate code, could perhaps do this on behalf of users even when they have limited coding skills (or none).
In all of these methods, there needs to be some kind of “firewall” at the gate of the interfaces that protects them from applications behaving badly. With APIs there is always a risk that someone on accident or purpose creates automations that behave badly. For example, send inordinate amounts of requests or go to eternal loops flipping something on and off endlessly. The interface needs to be able to detect these and do request throttling and inform the developer of their mistake.
ML-Me
In addition to creating composite services with a variety of static methods, today the use of machine learning in service creation is an obvious choice. And these services could be totally new kinds. What could that mean? Let’s take an example.
What if I would let artificial intelligence (AI) guide my life for a few weeks or a month or so. I would be like a vacuum cleaning robot where AI tells where to go next and what to do at every step:
When to wake?
What to eat for breakfast?
When should I go to bed and how long to sleep?
What clothes to put on in the morning?
When to leave for work?
What route to take?
Tell me what is my primary task for the day – i.e., what should I have done by the end of it?
Which emails to read and what to ignore, answer the simple ones for me, thank you very much
Accept the right invites via email and reject the requests I do not need to attend
When to leave work and route to take
Reminds what to do in the evening (shop on way to home and what to buy, or just shop for me as you already know what I want/ought to eat)
What to watch on TV, streaming services or not to watch?
What exercise to take or what book to buy or loan?
What to read/study?
Who to take out on a date?
What to plan for holidays?
If there is extra money, where to invest it?
And so on.
At least two big concerns raise from this. Does this reduce or destroy my humanity and my personality when computers start deciding? How could an inhuman machine take over my life and ever understand what I really need?
The best characteristic of such a system would certainly be that I can turn it off.
Or can I? What if I become too accustomed aka dependent on it?
The other much bigger worry is that what if my life would be much, much better, more enjoyable, healthy, varied and fulfilling this way. After all the experiences of a billion lives shared and combined create a much greater insight into what a good life is than my own fumblings and small experiments so far. Being outsmarted by the tiniest silicon chip instructing me in every step whether to turn left or right and when to turn my mental vacuum motor up or down.
What would be the background logic of this tiny silicon chip be? What if it is just optimising the bacterial contents of my gastrointestinal tract and this guarantees my happiness better than anything else I’ve done so far?
Would I ever want to know this?
Exercise for the gentle reader: use online search to find out more on gut-brain link and how it impacts our emotions. And for the advanced practitioners: try to find out how the microbial contents of our gut affects our political position and what party we vote for in elections.
And to that happy notion, it is time to end this little post.