Eric Drexler revisits his 2019 report, «Reframing Superintelligence,» in light of recent advances in Large Language Models (LLMs). He argues that his original model of AI development, which emphasizes AI services over unitary agents, remains relevant. The report advocates for a broader perspective on superintelligence, encompassing AI systems that are structured, transparent, and manageable. Drexler suggests that focusing on AI services can mitigate risks associated with superintelligent agents and open up new solutions for AI alignment, ultimately arguing that misaligned humans pose a greater threat than AI itself.
What is the main idea of «Reframing Superintelligence»?
Expanding the ontology of superintelligence; includes AI services; considers structures and relationships.
This perspective leads to the Comprehensive AI Services (CAIS) model, viewing general intelligence as flexible systems where task-focused agents are components.
What is the Comprehensive AI Services (CAIS) model?
General intelligence as flexible systems; task-focused agents; expanding AI services.
What is a key difference between AI services and AI agents?
AI services are task-focused; AI agents are unitary; AI services offer safety affordances.
Artículo Original: https://www.alignmentforum.org/posts/LxNwBNxXktvzAko65/reframing-superintelligence-llms-4-years
Advices:
- Focus on developing AI services rather than unitary AI agents to mitigate risks and improve manageability.
- Prioritize AI safety research that explores structured system development and safety guidelines for AI systems.
- Explore AI R&D automation to decouple AI improvement from AI agency, focusing on expanding safe AI functionality.
Deja una respuesta