Opinion piece regarding: General Outline Political Agreements and Concessions Within the To-Be House of Representatives in The Netherlands Regarding Major Subjects (Dutch: Hoofdlijnenakkoord Tweede Kamer)
Publishing Date: 02-06-2024 v1.0 / Author: Danile Dyadko \ Update Date: 03-06-2024 v1.1 (fixed html bugs)- General Note: Considering this is about politics in The Netherlands, some words are difficult to translate while maintaining their “true” meaning. I am aware that this may reduce reading pleasure but considering the highly interconnected world we live in, I made a conscious decision to formulate my opinions and views in English. Please rest assured that I have tried my best to translate any jargon to English to the best of my abilities. Another decision I made is to (almost) always include the original Dutch word in order to prevent miscommunication to be created. -
On Thursday 16th of May at 02:07, The most "realistic" (many variables are at play here) coalition of the to-be Dutch House of Representatives presented their “Hoofdlijnenakkoord” (for translation see title) to the Chairman of the Chamber. 2 hours and 7 minutes after the official deadline. The result is found here.
It does not come as a surprise to many Dutch citizens that the “Hoofdlijnenakkoord” took until the very last day to create considering the extreme views taken by some political parties and the general political distance between the political parties and their respective representatives. This was already noticeable from the last days of the electoral campaign last year of the House of the Representatives electorals where they would (quite rude sometimes) “bully” each other on national television . A general trend (in my opinion since the corona pandemic) towards disconnect between political parties and kindness can be observed and it’s trickle down effects on society Quite tragic but perhaps the unintentional effects of looking too much to America for our entertainment and secretly our politics. America is after all the leader in technology, GDP , entertainment, military, and a few other areas. New York used to be a part of The Netherlands so in light of the ever-rising nationalism and patriotism in The Netherlands, Americanizing politics may not come as much of a surprise after all.
What is interesting to note is that the “Hoofdlijnenakkoord” is only officially available in the Dutch language. Very likely this is unintentional (occam’s razor) but perhaps this is a slight nod to waters this potentially to-be coalitions want to fare in – less immigration, less international student, less European interference in some cases etc. (note: greatly simplified).
What caught my attention the most is the statement in Chapter 7 Statement 17 (Page 18):
“Er komt een wetenschappelijke standaard voor het gebruik van modellen en algoritmes. Beide moeten openbaar en navolgbaar zijn. De bijsluiter maakt duidelijk waarvoor ze wel en waarvoor ze niet bedoeld zijn en gebruikt kunnen worden.” (Unofficial translation: “There will be a scientific standard for the use of models and algorithms. Both must be public and understandable. The accompanying information will clarify what they are and are not intended for and can be used for.”)
After thinking quite some time about this statement since it’s publication (17 days ago) I still have the following thoughts that I have not found a satisfactory answers to:
1. “There will be a scientific standard …”
What entity will be responsible for creating/maintaining this scientific standard? If this entity is just The Netherlands, The Netherlands will be hindering their A.I. progress due to increasing bureaucracy. IT Documentation is not only a difficult aspect in IT to create and maintain (of IT businesses of all sizes) but fighting for its importance usually feels like an uphill battle (from all sides). Clients are more interested why something works than how it works. How can we blame them? Eventually they pay us to provide a service. For example: We apply this same logic (albeit largely unconsciously) to all things we use day to day. Does an average person know/cares how a train works on a low-granularity level? Does an average person know where their food comes from? Quite logical considering we can’t be questioning everything (considering this will go against our (neuro) biology of which we mainly use system 1 rather than system 2 (quick thinking 98% of the time)) because that would slow us down too much. Once again: this is something that is not in The Netherlands best interest to pursue.
Adding to the argument: the following official document was published: “Strategisch Actieplan voor Artificiële Intelligentie” (“Strategic Actionplan for Artificial Intelligence”) in 2019. We can see the following on page 10: “Urgentie om te versnellen. Nederland moet gaan versnellen op het gebied van AI. AI is namelijk cruciaal voor de efficiëntie en effectiviteit van alle sectoren en domeinen.” (“Urgency to increase speed. The Netherlands should increase speed in the A.I. industry. A.I. is crucial for the efficiency and effectiveness of all sectors and domains)
If it will be an European Agreement: What will prevent other (rising) global powers (BRIC) to ignore this agreement and instead of replicating a similar agreement – will ignore everything and value innovation (competitiveness/power) over anything else?
Quite strange if the to-be coalition will continue heading towards this direction considering the general anti-EU sentiment which is proven by multiple clauses within the “Hoofdlijnenakkoord”. One of such examples is to be found in Chapter 4 Point 2: "In Europa wordt met lef alles op alles gezet om Europese richtlijnen zo aan te passen dat ze werkbaar zijn en het verdienmodel ondersteunen" ("In Europe, bold efforts are made to adjust European guidelines in such a way that they are workable and support the business model."). This signals a distrust in the existing European guidelines that are in place to in-the-end increase prosperity for all it's citizens.
If it will be an agreement with every single country in the world, that would quite possibly one of the greatest geopolitical achievements of our time. Something that would come close to something similar would be the world-wide ban on ozone harming gasses in 1985 by which to current estimates the hole in the ozone layer will be sealed by 2045, 30 years earlier than previously predicted
Perhaps one of the most beautiful aspects of IT is that it is (to a certain degree, protocols etc. left out) free. You can choose your own programming language to your liking or make one yourself! You can choose your own code editor, your own specialisation and much more than I am not mentioning it right now. If I may quote an classmate “It feels as if I’m in a sandbox where I can make my own world”. I truly believe freedom increases innovation, not stifles it (regarding all endevours ever).
2. I have a sense that the word “algorithm” is used as a buzzword considering something as simple as mutiplication could be considered an algorithm. Who will define what an algorithm is?
3. All models and algorithms used in business and research should be public? I think it is clear that quite possibly everybody in current society would agree that A.I. would be to some degree beneficial to the progress of a society. Wouldn’t the public requirement force everybody to “out” secrets regarding A.I.? The Netherlands is already behind in this technology (name one Dutch innovative a.i. company non-consultancy related if you think this statement is false) and would be more so if this would be implemented. If this ideal is pursued, I think The Netherlands will act (against their will) as a backseat player in the next revolution of moving from mind work to creative work. Taking into account the fact that the financial landscape of The Netherlands will change too due to the three points (in my observation): (Forced) reduction of farming due to elevated levels of nitrogen in the ground, Halting of fracking gas extraction in our Province Groningen and general acceptance of the citizens, businesses and the government to (heavily) invest in energy transition with no to little financial benefit to show for it.
4. Intended information that would be mandatory to provide with the spread of models and algorithms would be formed in a way like the general terms of condition everybody just click the check mark box to and not read it. If it would be legally binding to not use certain models/algorithms for certain purposes, we would strain the National and International Security Team capacity even more. Considering there are far greater and very serious subjects that need to be addressed in cyberspace regarding security and law this would be considered an inefficient allocation of resources and even dangerous proposition. Considering software developers and especially software testers/auditors are not exactly high-supply jobs/skills It would be interesting to see how this would roll out. Perhaps implement an A.I. model to mandatory run withing Dutch IT Architecture would be a relative quick and easy solution.
5. The word traceable is used in the translation of the original word “navolgbaar” but perhaps a better (rough) translation that encapsulates the essence of the word would be: “understandable by humans”. This is by far the most tragic thing I have read in the whole “Hoofdlijnenakkoord”. There are a lot of A.I. models that are so called “black box”. You put information in, you get information out but what happens in-between could be either A. a big mystery or B. a slightly less big mystery. Considering some specialised A.I. models can (almost) always outperform humans (for example: AlphaGo (Great watch. If you have watched it, please share your thoughts to me through LinkedIn). How do you feel about it? Are you scared and filled with dread or are you fascinated and hopeful? I think whichever camp is the biggest will write the direction of humankind will go (black swan events left out of the equation of course).
I think it’s very egotistical (but frankly, quite logical considering we always view the world from our own perspectives) to believe we are the smartest species on earth and always will be. To take a simple example: Humans can only observe a small fraction of colours (radiaton) available in the world (0.0035% to be exact) . Are we the entities that can observe the most colour? No. Then why do we take such a defensive stance (albeit unconsciously) regarding morality and intelligence? I think we have to think very carefully of the input and architecture and output of A.I. models (and algorithms) but we also have to let the control go in a sense as a parent that lets their child bicycle on their own without sidewheels for the first time. Are we ready to take the sidewheels off or are we going to always be scared?