An arms race for an artificial general intelligence would be detrimental for and even pose an existential threat to humanity if it results in an unfriendly AGI. In this paper, an all-pay contest model is developed to derive implications for public policy to avoid such an outcome. It is established that, in a winner-takes-all race, where players must invest in R&D, only the most competitive teams will participate. Thus, given the difficulty of AGI, the number of competing teams is unlikely (...) ever to be very large. It is also established that the intention of teams competing in an AGI race, as well as the possibility of an intermediate outcome, is important. The possibility of an intermediate prize will raise the probability of finding the dominant AGI application and, hence, will make public control more urgent. It is recommended that the danger of an unfriendly AGI can be reduced by taxing AI and using public procurement. This would reduce the pay-off of contestants, raise the amount of R&D needed to compete, and coordinate and incentivize co-operation. This will help to alleviate the control and political problems in AI. Future research is needed to elaborate the design of systems of public procurement of AI innovation and for appropriately adjusting the legal frameworks underpinning high-tech innovation, in particular dealing with patenting by AI. (shrink)

In recent years the understanding of the cognitive foundations of economic behavior has become increasingly important. This volume contains contributions from such leading scholars as Adam Brandenburger, Michael Bacharach and Patrick Suppes. It will be of great interest to academics and researchers involved in the field of economics and psychology as well as those interested in political economy more generally.

The paper presents a variation of the EMAIL Game, originally proposed byRubinstein (American Economic Review, 1989), in which coordination ofthe more rewarding-risky joint course of actions is shown to obtain, evenwhen the relevant game is, at most, ``mutual knowledge.'' In the exampleproposed, a mediator is introduced in such a way that two individualsare symmetrically informed, rather than asymmetrically as in Rubinstein,about the game chosen by nature. As long as the message failure probabilityis sufficiently low, with the upper bound being a (...) function of the gamepayoffs, conditional beliefs in the opponent's actions can allow playersto choose a more rewarding-risky action. The result suggests that, forefficient coordination to obtain, the length of interactive knowledge onthe game, possibly up to ``almost common knowledge,'' does not seem to bea major conceptual issue and that emphasis should be focused instead onthe communication protocol and an appropriate relationship between thereliability of communication channels and the payoffs at stake. (shrink)

In a stimulating paper, Piccione and Rubinstein (1997) argued how a decision maker could undertake dynamically inconsistent choices when, in an extensive form decision problem, she has a particular type of imperfect recall named absentmindedness. Such memory limitation obtains whenever information sets include decision histories along the same decision path. Starting from work focusing on the absentminded driver example, and independently developed by Segal (2000) and Dimitri (1999), the main theorem of this article provides a general result of dynamically consistent (...) choices, valid for a large class of finite extensive form decision problems without nature. (shrink)

Time discounting is the phenomenon that a desired result in the future is perceived as less valuable than the same result now. Economic theories can take this psychological fact into account in several ways. In the economic literature the most widely used type of additive time discounting is exponential discounting. In exponential discounting, the fall of valuation depends by a constant factor on the length of the delay period. It is well known, however, that exponential time discounting often does not (...) describe well how people actually behave. Most people are averse to short delays in gratification now, while their future selves may not mind a bit of extra waiting. This behaviour can be described well by non-exponential discounting functions such as hyperbolic discounting. In hyperbolic discounting, valuations fall rapidly for small delay periods, but the fall gets slower for longer delay periods. Hyperbolic discounting captures phenomena such as procrastination, addiction and in general inconsistency over time. This chapter investigates whether forms of non-exponential discounting, in particular close to the so called Quasi-Hyperbolic model, could also be characterized in terms of dynamically consistent choices when individuals discount the welfare of future selves as well as their payoffs. (shrink)