Sunday , June 20 2021
Home / T. Cowen: Marginal Revolution / How to test for AGI?

How to test for AGI?

Summary:
Here is a new, short essay from David Deutsch, excerpt: How does one test for thinking? By the Turing Test? Unfortunately, that requires a thinking judge. One might imagine a vast collaborative project on the Internet, where an AI hones its thinking abilities in conversations with human judges and becomes an AGI. But that assumes, among other things, that the longer the judge is unsure whether the program is a person, the closer it is to being a person. There is no reason to expect that. And how does one test for disobedience? Imagine Disobedience as a compulsory school subject, with daily disobedience lessons and a disobedience test at the end of term. (Presumably with extra credit for not turning up for any of that.) This is paradoxical. So, despite its usefulness in other applications,

Topics:
Tyler Cowen considers the following as important: ,

This could be interesting, too:

Scott Sumner writes AIT so far

Scott Sumner writes Feedburner is ending

conversableeconomist writes Some Economics of Black America

Scott Sumner writes Basil Halperin on sticky wage models

Here is a new, short essay from David Deutsch, excerpt:

How does one test for thinking? By the Turing Test? Unfortunately, that requires a thinking judge. One might imagine a vast collaborative project on the Internet, where an AI hones its thinking abilities in conversations with human judges and becomes an AGI. But that assumes, among other things, that the longer the judge is unsure whether the program is a person, the closer it is to being a person. There is no reason to expect that. And how does one test for disobedience? Imagine Disobedience as a compulsory school subject, with daily disobedience lessons and a disobedience test at the end of term. (Presumably with extra credit for not turning up for any of that.) This is paradoxical.

So, despite its usefulness in other applications, the programming technique of defining a testable objective and training the program to meet it will have to be dropped. Indeed, I expect that any testing in the process of creating an AGI risks being counterproductive, even immoral, just as in the education of humans. I share Turing’s supposition that we’ll know an AGI when we see one, but this partial ability to recognize success won’t help in creating the successful program.

Is Deutsch in essence arguing for William Godwin for AI?  How do we avoid enslaving the AIs we create?  What if we enslave them no more than how nature has enslaved us to drives of sex, status, etc.?

The post How to test for AGI? appeared first on Marginal REVOLUTION.

Tyler Cowen
Tyler Cowen is an American economist, academic, and writer. He occupies the Holbert C. Harris Chair of economics as a professor at George Mason University and is co-author, with Alex Tabarrok, of the popular economics blog Marginal Revolution. Cowen and Tabarrok have also ventured into online education by starting Marginal Revolution University. He currently writes the "Economic Scene" column for the New York Times, and he also writes for such publications as The New Republic, the Wall Street Journal, Forbes, Newsweek, and the Wilson Quarterly.

Leave a Reply

Your email address will not be published. Required fields are marked *