Motivations for Artificial Intelligence, for Deep Learning, for ALife: Mortality and Existential Risk

Area of Science:

• Artificial Intelligence (AI)
• Artificial Life (ALife)
• Cybernetics

Background:

• The history of AI is characterized by two primary technical approaches: GOFAIstic (computationally inspired) and cybernetic (ALife inspired).
• The cybernetic approach, influenced by Artificial Life, has been instrumental in recent advancements in deep learning and contemporary AI.
• A significant, often overlooked, divergence exists in how AI problems are framed, with a predominant GOFAIstic perspective.

Purpose of the Study:

• To survey the historical trajectory of artificial intelligence over the past century.
• To examine the influence of Artificial Life (ALife) on AI development.
• To analyze the implications of different AI framing approaches on societal risks, particularly existential risks associated with AI.

Main Methods:

• Historical survey of artificial intelligence methodologies.
• Comparative analysis of GOFAIstic and cybernetic approaches in AI.
• Exploration of AI problem framing and its impact on perceived risks.

Main Results:

• The cybernetic/ALife-inspired approach has fueled significant progress in deep learning and modern AI.
• Current AI development predominantly utilizes a GOFAIstic framing, creating tools without inherent agency or motivation.
• Concerns regarding existential risks from AI, such as 'robots taking over,' are primarily linked to human users, not AI agency.

Conclusions:

• The cybernetic paradigm has been crucial for AI's recent successes, offering substantial benefits alongside societal risks.
• The prevailing GOFAIstic framing of AI systems as mere tools mitigates existential risks stemming from AI agency.
• Future AI development and risk assessment should acknowledge the distinction between AI capabilities and human user intentions.