The Risks of AI: An Interview with Georgetown’s Helen Toner

Journal of Political Risk, Vol. 10, No. 1, January 2022

Helen Toner, Director of Strategy at the Center for Security and Emerging Technology (CSET) at Georgetown University.

Anders Corr, Ph.D.
Publisher of the Journal of Political Risk

The JPR interview with Helen Toner, the Director of Strategy at the Center for Security and Emerging Technology (CSET) at Georgetown University, was conducted via email between 4 January 2022 and 13 January 2022.

Corr: What are the national security risks and benefits of AI?

Toner: This is a huge question! AI is a general-purpose technology, meaning that—like electricity or the computer—its impacts will be felt across practically all industries and areas of society. Accordingly, it presents a huge range of potential risks and benefits from a national security perspective. One way of trying to summarize the possibilities might be as follows: the benefits will largely be in line with the kinds of benefits we have seen from increasingly sophisticated computing technology more generally: greater efficiency and accuracy, as well as the ability to perform tasks at scales impossible for humans (think: how Google search trawls the web). In terms of risks, one breakdown proposed by Zwetsloot and Dafoe is to think in terms of risks from accidents (i.e. unintended outcomes from using AI), misuse (i.e. the deliberate use of AI to cause harm), and structural changes (i.e. how progress in AI shapes surrounding systems and dynamics). I realize this is fairly abstract, but it’s impossible to enumerate specific risks without narrowing the scope to particular application areas, time frames, and actors.

Continue reading

Legislatures Elected by Evaluative Proportional Representation (EPR): An Algorithm

Journal of Political Risk, Vol. 8, No. 1, January 2020

Source: Pixabay

Stephen Bosworth, Anders Corr, Stevan Leonard1

Abstract

Unlike existing voting methods, this article describes a new method that gives all voters every appropriate reason to be pleased. Evaluative Proportional Representation (EPR) invites each citizen to grade the suitability for office of any number of candidates as either Excellent (ideal), Very Good, Good, Acceptable, Poor, or “Reject” (completely unsuitable).  EPR allows each citizen to guarantee that one of the elected members of the legislature has received either their highest grade, remaining highest grade, or proxy vote – no vote is needlessly wasted. Continue reading

Legislatures Elected by Evaluative Proportional Representation (EPR): An Algorithm

Journal of Political Risk, Vol. 7, No. 8, August 2019

Stephen Bosworth, Anders Corr and Stevan Leonard1

Abstract

Source: Pixabay

Unlike any existing voting method for a representative democracy, this article describes a new method that gives every voter every appropriate reason to be pleased with the results. It is called Evaluative Proportional Representation (EPR). EPR guarantees that each citizen’s vote will continue to count proportionately in the deliberations of a legislative body, such as a city council. After assessing the ideal qualities needed by the office, citizens grade each candidate as either Excellent (ideal), Very Good, Good, Acceptable, Poor, or “Reject” (completely unsuitable). Each voter can give the same grade to more than one candidate. Each candidate not graded is automatically counted as a “Reject” by that voter. These grades can be counted by anyone who can add and subtract whole numbers or by the algorithm provided. Each EPR citizen’s vote adds proportionately to the voting power in the legislature of a winner. Initially, EPR’s count provisionally determines the number of highest grades (votes) each candidate has exclusively received from all the voters. However, no winner is allowed to retain enough votes to dictate to the legislature. Therefore, our simulated election limits the percent of votes any winner can retain to 20%. This ensures that at least three members of the legislature will have to agree for any majority decision to be made. We call a candidate who has received such a percentage super popular. Any non-super-popular candidate is eligible to receive at least one of the extra votes initially held by a super-popular candidate. Each extra vote is transferred to the remaining eligible candidate on this voter’s ballot who has been awarded the highest remaining grade of at least Acceptable. If such a candidate is absent, this ballot becomes a proxy vote that must be publicly transferred to an eligible winner judged most fit for office by this super-popular candidate. Similarly, all the votes provisionally held by an unelected candidate must be transferred to an eligible winner. The final number of votes received by each winner is the weighted vote each will use during the deliberations of the legislature. No vote is needlessly wasted. Each citizen is given every appropriate reason to be pleased.

Continue reading

Legislatures Elected by Evaluative Proportional Representation (EPR): An Algorithm

Journal of Political Risk, Vol. 7, No. 6, June 2018 

Illustration of grading. Source: Pexels.

Steve Bosworth and Anders Corr1

Abstract

This article describes a new and relatively simple evaluative method to elect all the members in any legislative body, such as a city council or national legislature.2 Called Evaluative Proportional Representation (EPR), each voter grades any number of candidates on their fitness for office as EXCELLENT, VERY GOOD, GOOD, ACCEPTABLE, POOR, or REJECT.  These evaluations are counted by hand or computer algorithm (here provided in the R statistical computer language).  This evaluative method of social choice is particularly good at revealing and optimizing voters’ utilities.  It ensures proportionate minority representation in legislative bodies by enabling each voter to guarantee that his or her evaluations of the candidates will continue fully to count in the deliberations and decisions made by their elected legislative body.  Each elected member of this body is given a different weighted vote as determined by counting all voters’ evaluations. As a result, each citizen’s vote continues to count within the weighted vote given to the elected member she most highly values.

Note: An updated version of this article is available at: https://www.jpolrisk.com/legislatures-elected-by-evaluative-proportional-representation-epr-an-algorithm-v2/

Continue reading