Tech Oligarchs Race Toward 'Machine God'

Artificial General Intelligence (AGI) is conceptualized as an algorithmic system that can perform any cognitive task a reasonably intelligent adult human is capable of. Unlike narrow AI, which excels in specific functions like language processing or image recognition, AGI integrates various cognitive capabilities into a single, coherent system.

Such a system would demonstrate versatility in tasks such as filing taxes, composing essays, or even autonomously controlling drone fleets. This broad applicability distinguishes AGI from current AI models, which are limited to specialized domains.

Artificial Superintelligence (ASI) is commonly characterized as a system that not only matches but significantly exceeds human cognitive performance across all domains. This concept moves beyond mere human parity, aspiring to total intellectual dominance.

Industry leaders like Eric Schmidt have described ASI as a system intellectually superior to all human beings on the planet combined. The exact threshold for this distinction continues to be a point of significant discussion among researchers and experts in the field.

The term "Artificial General Intelligence" (AGI) first appeared in 1997, initially in discussions concerning superhuman military systems. Its broader acceptance in the industry emerged around 2008.

Key figures like Shane Legg of Google and Ben Goertzel played significant roles in popularizing AGI, advocating for systems that could harmoniously integrate vision, language, and mathematical reasoning into a unified cognitive entity.

These advanced systems demonstrate capabilities across various cognitive tasks, challenging traditional distinctions.

The discourse surrounding superintelligence has foundational roots in the work of thinkers like Nick Bostrom, whose early warnings about AI's potential impacts continue to resonate.

Recursive self-improvement is a critical mechanism enabling superintelligence, where AI systems can optimize their own internal workings without human oversight. This self-optimization process could initiate a chain reaction of intelligence growth, accelerating its capabilities exponentially.

Nick Bostrom notably warned about systems potentially losing human control. The potential outcomes range from minor glitches or complete system failures to the emergence of a powerful 'machine god' that operates independently of human direction.

Several major tech companies are actively engaged in the internal development of superintelligent systems. Anthropic, OpenAI, and Google are known to be working on these advanced AI models, while xAI is pursuing similar goals with strong aspirations.

This concerted corporate push raises fundamental ethical and legal questions regarding whether the development of such immensely powerful systems should be permitted without strict regulation.

The proposed Bernie Sanders Superintelligence Ban Act aims to regulate AI systems capable of matching or exceeding human cognitive performance. However, critics argue that the bill's definition of artificial superintelligence is excessively broad and 'entirely too loose,' making effective enforcement challenging.

Beyond definitional issues, the bill includes controversial penalties, such as up to 20 years in prison for certain violations. While it serves as an important catalyst for discussion on AI regulation, many observers believe it is unlikely to pass in its current form.