Sign In

Communications of the ACM

ACM TechNews

Animal-AI Olympics Will Test AI on Intelligence Tasks Designed for Crows, Chimps


View as: Print Mobile App Share: Send by email Share on reddit Share on StumbleUpon Share on Hacker News Share on Tweeter Share on Facebook
The project marks a new way to evaluate the progress of AI systems toward what researchers call "artificial general intelligence."

The Animal-AI Olympics, which will begin in June, was created to benchmark the current level of various AIs against different animal species using a range of established animal cognition tasks.

Credit: Squidoodle

The University of Cambridge's Leverhulme Center for the Future of Intelligence and GoodAI, a research institution based in the Czech Republic, are partnering to launch the Animal-AI Olympics, which aims to benchmark the current level of various artificial intelligence (AI) systems against different animal species using a range of established animal cognition tests.

The project marks a new way to evaluate the progress of AI systems toward what researchers call "artificial general intelligence."

While AI systems have surpassed humans in a range of challenging competitions, including the board game Go, the poker game Texas Hold'em, and the video game StarCraft, these only demonstrate that AIs were very good at those specific tasks.

AIs have yet to demonstrate the kind of flexible intelligence that enables humans to reason, plan, and act in many different situations.

From IEEE Spectrum
View Full Article

 

Abstracts Copyright © 2019 SmithBucklin, Washington, DC, USA


 

No entries found