Futurism

AI: The Central Planning Fallacy on Steroids?

Members of the central planning committee visited Jozsi in his village, wearing their nice, long, brown leather coats. 

“How many piglets will be born, Józsi?”

“Well… You know, no one can tell.”

“Shut up, you dirty peasant! Tell us the number! How many piglets will be born?”

“Well, I don’t kn…”

Bang, they slapped Józsi on the face.

“Oh my, my… How many does it have to be?” cried Józsi. “How many is the Plan, comrades?”

“14.”

“Then that’s how many will be born. But… did you tell the pig?”

*

Thus started the joke of Hofi, the only comedian who was permitted to crack such jokes in communist Hungary. (For the rest of the joke scroll down to the end of the post.) According to the late communist regime, jokes like this could serve as a pressure valve to let the steam out. And why was there so much steam? Because the population had to endure the hardship imposed by communist central planning. If they could crack (pre-approved) jokes about it, they might be able to endure it in a slightly better mood. The nonsensical nature of central planning also featured in films made at that time.

The joke about the piglets told in layman’s terms that the 5-Year Plan was arbitrary, it prescribed things that were not even doable, perhaps not even relevant, just quantifiable, and that it incentivized (if not forced) everyone to report false statistics. Based on which the central planners made awful decisions on behalf of the entire country. But would those decisions be better if they knew about the exact number of piglets?

Sadly, no. But this is the fallacy many commit when they welcome AI as the new central planning authority. They admit that central planners of yesteryear were limited, but not that central planning itself was wrong. They compare modern-day surveillance and digital data-gathering to greasy papers and pencils that Józsi, the peasant used to report the number of piglets and conclude that AI is superior in collecting data. As of the decisions based on that data, AI-planning proponents still harbor a completely unfunded assumption that 1) it will be benevolent and 2) that this time it will be different.

With the emergence of artificial intelligence, hopes that things (starting with the economy) can get dictated from the top down resurfaced with renewed fervor and justification. The age-old fallacy that central planning in superior to individual planning and decentralized coordination now has the glorified excuse of machines sweeping through often the same data as overconfident economists and politicians always had.

The first mistake central planners make is mistaking data with knowledge.

Central planners only measure things they can quantify – and set goals accordingly: pigs, horses, and metric tons of coal and wheat. But that is not a proper description of the economy. Statistics is only a placeholder for knowledge. Counting everything we can is as close as any government can get to the god-like, all-knowing image they project about themselves. But statistics are a very poor measure of success, even before they are manipulated. And politically relevant numbers are always manipulated.

The second mistake is mistaking machine-collected data with unbiased knowledge. Another, more obvious mistake stems from the nature of statistics, starting with poor collection methods, misunderstanding what the data represent and what it doesn’t, and cumulating in the politically motivated manipulation of any set of statistics under political attention. Goodhart’s law asserts that whenever a statistical measure becomes a political target, it ceases to be a useful measure. The wrong incentives apply not only those who collect the data, but also those who handle the data. They use the poor input, fire up the statistical fog machine, all in order to get published, to get quoted, to get elected, to appease their superiors with easy-looking solutions, etc.

In short, people count what they can, look for what they want to find, and find what they want in any data. And that is before political oppression, such as socialism or nationalist central planning comes into play and incentivizes people such as Józsi to muddle the numbers.

Proponents of the ultimate central planning by AI would naturally dismiss the potential of poor source data because machines soak up metadata, not survey results.

They are also dismissive of the possibility of meddling with data, because AI supposedly collects data we didn’t know we create. For now.

But AI will also rely on already existing data – and if anyone had even the slightest business with government statistics, they should know that they are not only politically motivated (even in civilized countries), but they even change retroactively. And finally, proponents of machine-omnipotence keep forgetting that data is a commodity, has owner and price and access to it is selective.

Machines also have limitations in collecting data regarding the offline world – and if privacy is to be taken seriously, they should have even more of those limitations. And finally, data-crunching AI policymakers will have to deal with the problem of access – data that is owned by someone else, data that is not sold or too expensive.

A more effective enforcement of planning may also prove to be bad for its own survival, bringing on disasters much faster than incompetent central planning comrades ever could in countries oppressed under the mirage of communist central planning as citizens enforced reality and logic in any small and illegal means they could. Individuals struggling to make do under oppression in order to survive sadly also prolong that oppression.

The central planning fallacy is deeply rooted in the authoritarian mind.

Six characteristics particularly expose authoritarian thinkers to resort to central planning approach:

1) The locus of identity resting outside of the authoritarian thinker’s mind.

It is often vested in the point of view of the leader, while mentally dehumanize the populations under leadership. Whenever we discuss politics and politics, we all think and speak like tiny gods, hovering above society and proposing fixes from the god’s eyes view. We all think with the political leader’s head and propose what we would do in his place. Ban this, tax that. Dish out sticks and carrots, incentives and punishment to peasants (such as ourselves).

We don’t just dehumanize each other in the process, but sinfully simplify the world. And as a consequence of this thinking method, we end up allowing central planners to meddle in our lives with the same simplistic tools and condescending considerations.

An AI, no matter what kind of unlimited surveillance data it gets access to, will just be the same, simplifying force on humanity. And of course, it won’t do it by itself, there is no such think as a machine obtaining human-like will to power. Politicians will use it to further their own goals: keeping power.

2) Authoritarian thinkers subscribe to a static world view, most prominently in the form of having a desirable end game in mind.

The means to and end fallacy comes to authoritarian thinkers naturally. We all have a view of how the world should be like (the end), but not all commit the fallacy that any means are okay as long as they lead to that glorious end state. One of the most spectacular ends in intellectual history was the Socialist Revolution put forth by Marx, no wonder it lead to the most damaging and destructive central planning experiment in human history: communism. But if you now think that only self-proclaimed lefties have this mind bug, you must be disappointed. The vision of an “all-christian Europe” or a “homeowner society” are also cases in point, to name just two.

3) Authoritarian thinkers may be loud and belligerent, but the reason they so forcefully demand a strongman is the underlying sensation of feeling helpless regarding their own life.

Central planning lends itself as a (false) solution to this problem. People unthinkingly try to use the facilities of the state to influence each other, they are therefore exposed to any ruler who promises to take care of things. Whether by meddling in other people’s wallets or life choices – it is only a matter of taste. Ideology is only a superficial justification for this underlying behavior.

4) Authoritarian thinkers are also prone to covet some type of homogenization of society,

…either by economic strength or by lifestyle. Nothing says homogenization like a central planning authority that massages the population until we are all the way we are supposed to be. Conservative meddling with other peoples’ love lives and socialists’ meddling in other peoples’ wallets are equally distasteful and authoritarian and neither stands for individual liberty. From the viewpoint of citizen empowerment and control over our own lives, they are both equally damaging. The justifications are just a beauty patch on a big, nasty, central planning effort and they never hold water.

5) Authoritarian thinkers are fond of order and status at the detriment of opportunity, and nothing says hierarchy like a central organizing force. Actually, nothing else says so.

6) Authoritarian thinkers are all trying – one way or another – to separate choice from responsibility.

Either by taking away choices but making people bear the consequences (what you call ‘right wing’), or by leaving choices wide open, sometimes beyond the scope of possibility, but letting people get away without the consequences (what you call ‘left wing’).

Banksters gamble and found a way to socialize losses, basic income believers also want want all the choices but not bear the economic consequences. Fundamentalist societies deny most choices in private life while leave people to cope with the consequences – like an unhappy life, an abusive marriage or unwanted parenthood -alone. Their opposition wants people to have all the choices (today a chef, tomorrow an artist), but refuse to assign individual responsibility for such choices. The examples are endless.

Central planning is just a very complex way of achieving the same, it promises whichever wonderland you wish to inhabit: the endless choices without responsibility, or the lack of choices but full responsibility.

These thinking characteristics make it all but impossible not to make an individual adopt the central planning approach. And AI will be just as politically motivated as the humans who wield it in their own interest – only perhaps better at that. When it comes to recognizing and allowing diverse strategies, trials and failures, innovation and non-typical life strategies, AI can be only slightly better than humans. Knowing which butterfly wing to flap in order to affect laser-sharp change exactly where it is deemed necessary (by whom, according to what priority?) is still out of reach. But the humans who can’t want to outsource decisions to that are already hot on the promise of it. It tells more about them than it tells about AI.

*

The joke of Hofi ended on a less than amusing note – even though people laughed at it out of helpless desperation.

As time passed, the piglets were born. But the pig birthed only 10 piglets. The Party secretary was frightened:

“Dear Holy Mary, what shall we do now? Only 10 piglets were born, but the Plan was 14. It may even be called a sabotage, what will happen to us?”

“I can’t make piglets” said Józsi. “But I can make statistics – Hungarian style! I’ll report 11 piglets. It’s not ten, after all. And it’s almost 14!” 

The paper with the statistics moved up to the commune level. 

“It’s outrageous, comrades! We can’t do this to the workers of the commune! So we will report 12 piglets.”

12 piglets it is, the Plan is going according to plan, the report arrives to the district.  

“12?” they said. “Comrades, that’s not enough. We will report 13.”

Report keeps moving up. 

“13, comrades? It makes me sad. Wasn’t there one more?”

So they report 14. The Plan is complete. Long live Comrade Rákosi! 

Comrade Rákosi, the Party Secretary of the Hungarian Communist Party addresses the workers:  

“I am pleased to welcome the world class quality results of socialist production, the 14 piglets. So we have decided to export 10 out of the 14 piglets – and we eat the rest!”

Follow us on Facebook , Twitter @_MwBp , or our new Telegram channel @budapesthungary 

Featured image: Helicopter taxi in Moscow, 1956. From the journal “Knowledge is Strength”

 

 

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.