Thinking about a scenario with possible risk of flooding and landslides caused by weather conditions, it results really interesting to investigate the way in which citizens take decisions on the basis of different information sources they can access. In this work we start describing a platform we realized in order to study this social phenomenon. Then we present some simulative experiments showing how a population of cognitive agents trusting in a different way their information sources can make decisions more or less suited to the several weather patterns. The complexity of decisions is based on the fact that the agents differently trust the various sources of information, which in turn may be differently trustworthy. In our simulations we analyse some interesting case studies, with particular reference to social agents that need to wait others in order to make decision.

Due to Internet and social media web, the world as we know it is deeply changing integrating two different aspects of the social interaction: the one that develop in the real world and the one that develop in web society. In this paper we focus on the importance of generalized knowledge (agents' categories) in order to understand how much it is crucial in these two worlds. The cognitive advantage of generalized knowledge can be synthesized in this claim: "It allows us to know a lot about something/somebody we do not directly know". At a social level this means that I can know a lot of things on people that I never met; it is social "prejudice" with its good side and fundamental contribution to social exchange. In this study we will analyse and present some differences between the social relationships in the two worlds and how they influence categories' reputation. On this basis, we will experimentally inquire the role played by categories' reputation with respect to the reputation and opinion on single agents: when it is better to rely on the first ones and when are more reliable the second ones. We will consider these simulations for both the two kind of world, investigating how the parameters defining the specific environment (number of agents, their interactions, transfer of reputation, and so on) determine the use of categories' reputation and trying to understand how the role played by categories will be important in the new digital worlds.

In this article, we are interested in the fact that relevance and trustworthiness of information acquired by an agent X from a source F strictly depends and derives from X's trust in F with respect to the kind of information. In particular, we are interested in analyzing the relevance of F's category as indicator for its trustworthiness with respect to the specific informative goals of X. In this article, we analyze an interactive cognitive model for searching information in a world in which each agent can be considered as belonging to a specific agent's category. We also consider variability within the canonical categorical behavior and consequent influence on the trustworthiness of provided information. The introduced interactive cognitive model also allows evaluation of the trustworthiness of a source both on the basis of its category and on past direct experience with it, thus selecting the more adequate source with respect to the informative goals to achieve. We present a computational approach based on fuzzy sets and some selected simulation scenarios together with the discussion of their more interesting results.

In this article we consider the special abilities needed by agents for assessing trust based on inference and reasoning. We analyze the case in which it is possible to infer trust towards unknown counterparts by reasoning on abstract classes or categories of agents shaped in a concrete application domain. We present a scenario of interacting agents providing a computational model implementing different strategies to assess trust. Assuming a medical domain, categories, including both competencies and dispositions of possible trustees, are exploited to infer trust towards possibly unknown counterparts. The proposed approach for the cognitive assessment of trust relies on agents' abilities to analyze heterogeneous information sources along different dimensions. Trust is inferred based on specific observable properties (manifesta), namely explicitly readable signals indicating internal features (krypta) regulating agents' behavior and effectiveness on specific tasks. Simulative experiments evaluate the performance of trusting agents adopting different strategies to delegate tasks to possibly unknown trustees, while experimental results show the relevance of this kind of cognitive ability in the case of open multiagent systems.

Typical solutions for agents assessing trust relies on the circulation of information on the individual level, i.e. reputational images, subjective experiences, statistical analysis, etc. This work presents an alternative approach, inspired to the cognitive heuristics enabling humans to reason at a categorial level. The approach is envisaged as a crucial ability for agents in order to: (1) estimate trustworthiness of unknown trustees based on an ascribed membership to categories; (2) learn a series of emergent relations between trustees observable properties and their effective abilities to fulfill tasks in situated conditions. On such a basis, categorization is provided to recognize signs (Manifesta) through which hidden capabilities (Kripta) can be inferred. Learning is provided to refine reasoning attitudes needed to ascribe tasks to categories. A series of architectures combining categorization abilities, individual experiences and context awareness are evaluated and compared in simulated experiments.

Aim of this work is to provide a formal characterization of those emotions that deal with normative reasoning, such as shame and sense of guilt, to understand their relation with rational action and to ground their formalization on a cognitive science perspective. In order to do this we need to identify the factors that constitute the preconditions and trigger the reactions of shame and sense of guilt in cognitive agents, that is when agents feel ashamed or guilty and what agents do when they feel so. We will also investigate how agents can induce and silence these feelings in themselves, i.e. the analysis of defensive strategies they can employ. We will argue that agents do have control over their emotions and we will analyze some operations they can carry out on them.

Some symbolic AI models, for example, BDI (belief, desire, intention) models are conceived as explicit and operational models of the intentional pursuit and belief dynamics. The main concern of these models is to provide a clear understanding of the functional roles of different kinds of epistemic and motivational states (beliefs, acceptances, expectations, intentions, goals, desires, etc. …), of the relational properties among them. Mental configurations of appraisal (involving different kinds of motivational and epistemic states) which correspond to particular cognitive emotions such as disappointment, fear, relief, shame, etc. … have been analyzed by several authors close to the BDI theoretical tradition. The main objective of this work is a formal analysis of Surprise in a BDI-like cognitive architecture. A clarification of the functional role of Surprise in a BDI-like cognitive architecture with respect to resource bounded belief revision is given.

The general goal of the paper is to show the normative/deontic nature of conventions. Conventions are traditionally defined as regularity of behavior based on expectations evolved to solve coordination problems [14]. The thesis we defend is that the cognitive attitude of expectations is not only characterized by an anticipatory representation (belief) of a future state of affairs but is coupled with a motivational component (a goal on this state). The possible convergence between beliefs and corresponding goals allows the identification of positive and negative expectations. We argue that in positive expectations (differently from the negative ones) lies implicitly an influencing act aimed at prescribing that the expected event will be realized. We consider conventions as analyzed in Game Theory as regularity of behavior based on positive expectations. These conventions entail the deontic component of prescription. Each agent prescribes (and is subject to prescription) conformity to the convention to the others (prescription to do). This is a possible route to the spontaneous emergence of Social Norms. However we hypothesize, differently, that negative expectations too can sustain conventions. Even bad habits share a deontic component but is characterized by the socio-cognitive structure of permission (entitlement to do). We argue that with this analysis is possible to explain the self-organizing and stabilizing effect of conventions that create an equilibrium noxious for all the participants and individually more costly than the individual benefit.

In this paper, we describe how agents can deceive within a probabilistic framework for representing their mental state: in doing so, we challenge the so‐called sincerity assumption in Human–Computer interaction (HCI) and multi‐agent systems (MAS). We distinguish deception from its special case of lie and characterize different forms of deception, by identifying several criteria for distinguishing among them. In particular, we propose a model of information impact on the Receiver's mind. As the message Sender must plan its strategy by considering the Receiver's criteria for believing, we also discuss some of these criteria, like content plausibility, source informativity, and information safety. We apply this model to a simplified version of Turing's Imitation Game and describe how we implemented a Simulator of deceptive strategies that we called Mouth of Truth. We conclude the paper by describing an evaluation study that enabled us to verify the validity of our method and to revise it in part.

The relationship between trust and control is quite relevant both for the very notion of trust and for modelling and implementing trust-control relations with autonomous systems. We claim that control is antagonistic of the strict form of trust: 'trust in y '; but also that it completes and complements it for arriving to a global trust. In other words, putting control and guaranties is trust-building; it produces a sufficient trust, when trust in y 's autonomous willingness and competence would not be enough. We also argue that control requires new forms of trust: trust in the control itself or in the controller, trust in y as for being monitored and controlled, trust in possible authorities, etc. Finally, we show that paradoxically control could not be antagonistic of strict trust in y , but it can even create, increase it by making y more willing or more effective. In conclusion, depending on the circumstances, control makes y more reliable or less reliable; control can either decrease or increase trust. A good theory of trust cannot be complete without a theory of control.

In this paper, I claim that "information" is a social construct, is created and tailored on purpose by somebody for somebody else or collectively, accepted, believed, or propagated through social interactions. In particular, I argue about the essential social nature of several crucial aspects of information in IT and specifically in relation to Information Agents. After situating this view of information within the broader perspective of Social Artificial Intelligence — the new AI paradigm — I analyize the intrinsic social aspects of information ontology, search and access, presentation, overload, credibility and sources, value. The issue of real "collaboration" in providing information deserves special attention and I introduce our theory of over-help and of the ability to provide something different from what is requested in order to satisfy the real need beyond the request. In the second part, I focus on the crucial role of trust (an intrinsically social and cognitive notion) for dealing with information and sources especially on the web. My objective is to claim that: if information per se is a social construct, then its technology should be socially designed and integrated, and that machines must be involved in real social relationships because they have to mediate them among humans. It is necessary incorporating some part of this social knowledge and capability in the information technology itself, especially in adaptive and interacting "agents" and MAS.

We refer to social autonomy in a collaborative relationship among agents based on delegation and help. We address the problem of adjustable autonomy. We stress in particular the role played in autonomy by: 1) the degree of "openness" of delegation; 2) the allowed initiative in (re)starting negotiation; 3) the degree and kind of control; and 4) the strength of delegation with respect to interaction. We show how the adjustability of delegation and autonomy is actually "bilateral". Adjustment is also "bidirectional" and multidimensional. Finally, we analyze some reasons for modifying the assigned autonomy and show how the adjustment of autonomy depends: on the delegator's side; on a "crisis of trust" and vice versa; the delegee's adjustment of its own autonomy depends on some disagreement about the trust received from the delegator, and, in particular, either a higher or lower confidence in itself or in external circumstances. Some preliminary hints about necessary protocols for adjusting the interaction with agents are provided. This work is aimed at providing a theoretical framework, i.e., the conceptual instruments necessary for analyzing and understanding interaction with autonomous entities.

In this paper we will briefly present the adjustment of autonomy as the problem of adjusting the level and the kind of delegation. The cncept of autonomy here considered is about a delegated task (or role). It is argued that by a careful analysis of the various aspects of delegation (openness, strength, etc.) it is possible to represent and model many dimensions of autonomy and consequently both the reasons and the ways to introduce a dynamic level of control.

Formalisms based on a propositional representation demonstrate a high representational competence because of their flexibility, properties, and their great expression power, even if they often fail to represent the meaning of the nodes that make up the propositional networks. The presence of a network of definitions would make the representation richer, and would augment the expressiveness and the inferential capabilities of the whole system: the twofold architecture of the hybrid systems seems to be suitable to satisfy these prerequisites. Nevertheless, the hybrid systems fail to clearly define the intensional and extensional levels of knowledge. This paper proposes a different distribution of tasks between the two parts of a hybrid system on the ground of a precise distinction among intensional, mental extensional, and extensional aspects of knowledge. Moreover, it suggests employing a propositional approach also for arranging the abstract definitions of the terminological box. These principles have been tested in KRAM (Knowledge Representation for Agency Modelling), a hybrid system used as a component both of a cooperative interface and of a system able to model actions and minds of cognitive agents.