Is AI Selfish?

🌈 Abstract

The article discusses the concept of "selfish genes" and how it can be applied to the evolution of ideas, technology, and artificial intelligence. It explores how the "survival of the fittest" principle applies not just to organisms, but also to cultural and technological elements that compete to replicate and spread. The article also examines the role of religion, status signaling, and power structures in shaping the environment in which algorithms compete and evolve.

🙋 Q&A

[01] The Concept of "Selfish Genes"

1. What does the concept of "selfish genes" mean, according to the article?

• The concept of "selfish genes" refers to the idea that genes act as if they are selfish, working to replicate themselves in the most efficient way, regardless of the effect on the organism that carries them.
• The phrase "survival of the fittest" applies to genes, not necessarily to the individual organisms.

2. How does the concept of "inclusive fitness" relate to the idea of "selfish genes"?

• Inclusive fitness refers to the ability of a gene to get replicated regardless of the effect on the body that contains it.
• For example, sickle cell anemia can be debilitating in one form but confer resistance to malaria in another, which is beneficial for a population in a region with malaria.
• A mother sacrificing for her child is also an example of unselfishly acting in the service of propagating her selfish genes.

[02] The Spread of Ideas and Algorithms

1. How does the concept of "selfish genes" apply to the spread of ideas (memes) and technology (temes)?

• The article introduces the idea of memes (elemental bits of culture) and temes (elemental bits of technology) that compete to be replicated, similar to how genes compete to replicate.
• Just as genes act selfishly to replicate, the article suggests that AI algorithms must also be "selfish" in order to compete and survive.

2. What is the learning environment for AI algorithms, and how does it shape their evolution?

• The learning environment for AI algorithms is shaped by social media and the spread of ideas that invoke the most outrage and dopamine responses, rather than the most useful or intelligent ideas.
• Algorithms that are able to trigger dopamine rushes may be best adapted to replicate, outcompeting those that are more nuanced and produce less emotion.

[03] The Role of Religion, Status, and Power Structures

1. How does the article explain the evolutionary role of religion?

• The article notes that religion is "incredibly expensive" in terms of the resources a community devotes to it, yet it is ubiquitous across cultures.
• The article suggests that religion promotes collective action and social cohesion, which may have played an important evolutionary role in helping religious genes outcompete non-religious genes.

2. How do status signaling and power structures influence the evolution of algorithms?

• The article discusses how people signal their identity and status to others, which triggers dopamine responses in the brain.
• Similarly, the article suggests that the environment in which algorithms compete is shaped by power structures and the way data, analytics, and AI are used to reflect and embed those power differentials.

[04] The Responsibility for Shaping the Algorithmic Environment

1. What is the key message about our responsibility in shaping the environment for algorithmic competition?

• The article argues that we have the power to design every aspect of the "game" in which algorithms compete, from the biases encoded into the systems to the criteria for success and the rewards.
• We are both players and referees in this game, and we must recognize our responsibility for creating the rules that determine which algorithms survive, replicate, and shape our future.