A Definition and a Test for Human-Level Artificial Intelligence

TL;DR

This paper proposes a definition and a test for human-level AI based on the ability to learn from language without explicit rewards, emphasizing language acquisition as a key characteristic.

Contribution

It introduces a new classification of intelligence based on learning methods and proposes a novel test for HLAI centered on language acquisition without explicit rewards.

Findings

Language-based learning can serve as a sufficient test for HLAI

A classification scheme for different types of intelligence

Highlights the importance of language in human-level AI development

Abstract

Despite recent advances of AI research in many application-specific domains, we do not know how to build a human-level artificial intelligence (HLAI). We conjecture that learning from others' experience with the language is the essential characteristic that distinguishes human intelligence from the rest. Humans can update the action-value function with the verbal description as if they experience states, actions, and corresponding rewards sequences firsthand. In this paper, we present a classification of intelligence according to how individual agents learn and propose a definition and a test for HLAI. The main idea is that language acquisition without explicit rewards can be a sufficient test for HLAI.