Modern Governance Flaws

sturlabragason.github.io Modern Governance Flaws

Writing about some of the stuff I’ve been working on or thinking about, as time permits.

https://sturlabragason.github.io/blog/2023/08/20/Goverment.html

Comradeship // Freechat @lemmygrad.ml

sturlabragason.github.io /blog/2023/08/20/Goverment.html

3 0

You're viewing a single thread.

8 comments

Though the author addresses the flaws of the proposal of algocracy, he conviniently forgets the fault which is the downfall of all the will thought out plans. Human emotions, greed and ambition.
Just as democracy in authors words is nothing more than a beauty pageant. The destruction of democracy are those reasons, and will also be the destruction of algocracy.
Author also side steps philosophical questions regarding AI such as the trolley problem.
Until these problems are solved, algocracy might be even more destructive than what democracy is currently.
- You are correct, I did not address these, merely for the sake of brevity. I am acutely aware of human greed, it's a feature not a bug 😅. In fact, algocracy is the weapon of choice here because of exactly this.
  Algocracy aims to reduce the impact of human emotions, greed, and ambition. Calling democracy just a 'beauty pageant' underscores the search for a more stable system, one less affected by our natural flaws.
  The trolley problem isn't exclusive to AI. They challenge human judgment just as much, if not more. What algocracy offers is a consistent platform to tackle these dilemmas based on collective societal input. The proposal focuses on AI systems overseen by human direction, reinforced by open-source principles and transparency. Furthermore, its accessibility could be enhanced by allowing individuals to inquire about the system and propose enhancements through LLM interactions.
  Arguing that algocracy could be more destructive than current democratic practices is a hasty judgment. The vision is to refine and mold it, leveraging open-source methods and transparency, into the best version of governance we can muster. An iterative, ever-improving model, always under the watchful eyes of society.
  
  Thanks for the discussion.
  Trolley problem is not exclusive to AI but consequences of a dissociated program taking a decision is less acceptable as compared to a committee of humans. For all AI it would merely be an equation to solve with no inkling of what that consequence mean. As a human condemning another human to death is much more than a mere dilemma.
  The solution as you proposed is human intervention. The question arises however that who would choose those humans. A learning AI might learn to choose only those humans who would not oppose is decision. Any product of human intellect will have the same biases and faults as the humans those who have made them. There are no true creations but rather inspiration from who we are and what we see.
  Whereas human overseers chosen by us would have the same problems as it is with democracy.
  I am not against using ai but replacing humans is not and should not be the aim.
  
  I appreciate the insights you've shared. Here's a clarification on the core points:
  Algocracy vs. Human Emotion: The objective isn't to replace the entirety of human decision-making but to use algocracy in areas where certain human flaws can be systematically minimized. Algorithms, when designed properly, can reduce the impact of biases and inconsistencies.
  Selection of Human Overseers: The selection would remain transparent and lean towards open-source principles. Individuals would be chosen by peers and those deemed competent within the community. We migh utilize a meritocratic governance model combined with token-based voting for stake, enhanced by a reputation system to ensure decisions are made by competent, committed individuals, all underpinned by open-source principles for transparency and collective oversight (this is oversimplifying it, there is a huge amount of great ways; I hope to have time later to write about them).
  Purpose of Algocracy: The goal is not an AI or algorithmic "takeover." It's about integrating algorithmic governance in specific areas to achieve consistency and fairness, always with human oversight at its core.
  The emphasis is on refining our governance systems, not replacing the human touch. Challenges exist, but with collaboration and transparency, we can navigate them effectively.

8 comments