We clearly can’t trust humans to run the world. Now we’ve gotta try the AIs.
Did you catch up with the troubling book “How Democracies Die” by Harvard professors Steven Levitsky and Daniel Ziblatt? Hmm the august provenance is already going to be a turnoff for the people who need it most. Anyway, you get the idea. Clearly a key part in our current Zeitgeist.
The recent G20 meeting was held in the now-largest country, that just happens to be rapidly leaving the democratic column. China was already there and is doubling down. Ditto for many other countries that we formerly saw as being democratic such as Hungary and Poland.
It’s clear that democracy is on the way out. Even the US is slipping backwards, and there’s no guarantee it won’t go further. As Freedom House, one of the few remaining institutional defenders of freedom states, “In every region of the world, democracy is under attack by populist leaders and groups that reject pluralism and demand unchecked power to advance the particular interests of their supporters, usually at the expense of minorities and other perceived foes.”
The Russian mercenary Prigozhin’s recent death just reminds us that there are new mercenary groups constantly being launched to roll back any promising democratic seeds, such as we see in Africa and the Middle East. The anti-freedom mercenary movement has given the latest push to the rollback of democratic governance throughout the world. Add that to the global takeover of social media by anti-democratic forces and you have the makings of the perfect anti-freedom storm.
If one thing is becoming clear, it is that you can’t trust humans to safeguard freedom and democracy. Backsliding is too tempting and easy and even the best women and men do it too easily. There’s apparently no gene for freedom, even though we have it for so many other behaviors.
It seems like the best way forward is to assume that the behavioral and biological bases for freedom and democracy either don’t exist or are just too weak to survive. That means we have to assume from the get-go that humans are terrible subjects for launching and maintaining freedom-loving countries. Somehow, we have to remove humans from the equation. Over to the AIs.
Naturally you’ve got to put something else in their place, in the form of a framework for defining values, behaviors and rules of governance so that the AIs get at least the basic tuition they need to do their job. So far AI has been addressing the basic technological issues and is far away from doing anything like that. But there’s already companies working in it such as the little-known Anthropic which calls its approach “Constitutional AI”. Theres a long way to go of course, but its hard to see any other alternative.
Somehow, we have to re-engineer homo politicus into AI politicus and to build the AI constitution so that the two can work together. We could call the result electronic democracy but that understates the job that needs to be done.
AIs need to be designed, developed and refined that can cope with social, political and behavioral variables; simultaneously we need to incrementally re-engineer humans not so that they can do the job of creating and maintaining freedom since we know they can’t but to support the AIs do the job for them.
The idea is to re-engineer humans that can work in tandem with the AI politicus so as to achieve the ideal harmonious and productive political architecture. AI politicus is not designed to totally replace humans but to complement them and undertake the actual operations of leadership to keep the controls of governance away from humans’ sticky fingers that might be too tempted to take over again and botch up the whole enterprise (again).
This implies new approaches to biotech and medical science that can identify and develop the appropriate genetic and biochemical structures needed to build biological elements needed to make the right handshakes with AI politicus and constitutional AI.
In order for this to all work we need to work on behavioral and biological engineering so that the appropriate biological components are in place to work in tandem with constitutional AI.
Of course, there will be nay-sayers who assert that this is all impossible. We can never build humans, AIs and constitutions that could ever get to the level of sophistication needed to remove humans from their own governance. For starters the AIs would never be innovative enough to be able to make the components needed.
That’s being tested as we speak. Did you see the recent experiment that pitted AIs against MBA students at Wharton? The aim was to see which group was more innovative. It wasn’t just in terms of quantity of ideas, but quality and overall attractiveness.
Yep, the humans lost. By a mile.
I know there’s a bunch of objections. But you get the idea. You can’t assume that humans would always out-lead and out-think AIs. It may well be that the decline of democracy and freedom is the result of a natural experiment that ran itself to see to what extent human leadership has the gas for a more complex world of governance and has come up short.
If so, bottom line, it is telling us that humans had their chance and screwed it up and it’s time to change horses, even if they are artificial too.
When you subscribe to the blog, we will send you an e-mail when there are new updates on the site so you wouldn't miss them.