Many companies have been using chatbots to provide automatic, scalable, and personalized customer care services. You may think building a chatbot to this end is an easy task, just defining a sequence of messages according to the standard conversation flow, right? No, there is a lot more to it than that! In order to deliver a pleasant experience to end-users (or simply users) and to keep them engaged, the Conversational Designers need to put in quite a bit of effort. Usually, they deep-dive into diverse articles and studies to better understand the business.

This is the first part in a collection of articles covering IoT and AI. There is a parable in Buddhism about problems: It says that we all have 83 problems, and as we take one away, another appears to take its place. I don't know about you, but I am pretty certain that the same applies to the Pareto principle. Many are all familiar with the Pareto principle, or the 80/20 rule: 20% of your effort can yield 80% of the desired value, while 80% of your can yield 20% of the desired value. Whether it be sales, relationships, learning or the arts, this rule informs us that focusing our efforts on the right 20% will drive the most impactful outcomes.

With the rise of artificial intelligence, it's very likely that robots will soon be replacing some U.S. jobs. According to a new Brookings Institution report, a quarter of U.S. jobs will be severely disrupted as artificial intelligence (AI) accelerates the automation of existing work. The report explains how roughly 36 million Americans hold jobs with high exposure to automation. However, many economists have found that AI and automation have an overall positive effect on the labor market. Matias Cortes, an assistant professor at York University, explains in this interview at the Warwick Economic Summit how these trends can create economic growth, reduce prices, and increase demand while also creating new jobs that make up for those that disappear.

Now that machine learning algorithms lie at the center of many resource allocation pipelines, computer scientists have been unwittingly cast as partial social planners. Given this state of affairs, important questions follow. What is the relationship between fairness as defined by computer scientists and notions of social welfare? In this paper, we present a welfare-based analysis of classification and fairness regimes. We translate a loss minimization program into a social welfare maximization problem with a set of implied welfare weights on individuals and groups--weights that can be analyzed from a distribution justice lens. In the converse direction, we ask what the space of possible labelings is for a given dataset and hypothesis class. We provide an algorithm that answers this question with respect to linear hyperplanes in $\mathbb{R}^d$ that runs in $O(n^dd)$. Our main findings on the relationship between fairness criteria and welfare center on sensitivity analyses of fairness-constrained empirical risk minimization programs. We characterize the ranges of $\Delta \epsilon$ perturbations to a fairness parameter $\epsilon$ that yield better, worse, and neutral outcomes in utility for individuals and by extension, groups. We show that applying more strict fairness criteria that are codified as parity constraints, can worsen welfare outcomes for both groups. More generally, always preferring "more fair" classifiers does not abide by the Pareto Principle---a fundamental axiom of social choice theory and welfare economics. Recent work in machine learning has rallied around these notions of fairness as critical to ensuring that algorithmic systems do not have disparate negative impact on disadvantaged social groups. By showing that these constraints often fail to translate into improved outcomes for these groups, we cast doubt on their effectiveness as a means to ensure justice.

Manipulation and bribery have received much attention from the social choice community. We consider these concepts in the setting of preference formalisms, where the Pareto principle is used to assign to preference theories collections of optimal outcomes, rather than a single winning outcome as is common in social choice. We adapt the concepts of manipulation and bribery to this setting. We provide characterizations of situations when manipulation and bribery are possible. Assuming a particular logical formalism for expressing preferences, we establish the complexity of determining a possibility for manipulation or bribery. In all cases that do not in principle preclude a possibility of manipulation or bribery, our complexity results show that deciding whether manipulation or bribery are actually possible is computationally hard.

The scores can be used in themselv es or serve as the basis for ranking or choice. For the present, only a few scoring pro cedures are endowed with their axiomatic characterizations. At the same time, a large num ber of ingenious procedures are advocated and used in such disciplines as manageme nt science, operations research, psychometrics, applied statistics, processing of spor t tournaments, graph theory, etc. Very few social choice papers deal with them. The aim of this pa per is to take one circumspect step toward an axiomatic framework for comparin g the merits of these elaborate procedures. As a result, we would like to isolate a family of s coring procedures that comprises a majority of'reasonable' procedures (so that th e further axioms could be imposed on this family). Two main approaches are applicable. The first one is to express the desired properties axiomatically, the second is to gather the ex isting procedures and specify their common algebraic form.