2005 - The Future of AI - John McCarthy

Texto

The long-term goal of AI is human-level AI. This is still not directly definable, although we still know of human abilities that even the the best present programs on the fastest computers have not been able to emulate, such as playing master-level go and learning science from the Internet. Basic researchers in AI should measure their work as to the extent to which it advances this goal.

AI research should not be dominated by near-term applications. DARPA should recall the extent to which its applied goals were benefitted by basic research. NSF should not let itself be seduced by impatience. The so-called AI winter was dominated by people who lost money in companies. That’s a common phenomenon.

I think the best hope for human-level AI is logical AI, based on the formalizing of commonsense knowledge and reasoning in mathematical logic.

Formalizing common sense requires extensions to mathematical logic including nonmonotonic reasoning and extensive reification, e.g., of concepts and also contexts. The reifications require appropriate reflection schemas.

Approaches based on simulating aspects of the human nervous system, such as neural nets, may also work. Big changes in approach are almost certainly needed, e.g., systems that distinguish appearance from reality, such as learning about three-dimensional objects from two-dimensional images or from a sense of touch.

Enlaces

URL: https://ojs.aaai.org/index.php/aimagazine/article/view/1842

Wayback Machine: http://web.archive.org/web/20220503225431/https://ojs.aaai.org/index.php/aimagazine/article/view/1842