A (Persuasive?) Speech on Automated Persuasion

Philosophers of language have taught us that at the basis of language production there is the intention to change the state of the world by intervening linguistically on other agents. Persuasion, being the process of influencing attitudes, beliefs, behaviors, mood of a target, is a matter of stronger emphasis. Argumentation is just one resource to persuasion; it has been studied since the times of Aristotle and now for quite some time in artificial intelligence. The peripheral route to persuasion [1] is a different modality, one that is based on indirect, evocative, aesthetic aspects of the message. Automated intelligent persuasion of this sort (and also defense from inappropriate persuasion) is a research area close to producing usable results, both through creative production of language expressions, and through other forms of communication. The traditional goal of human-oriented information technology is mostly to offer services. With intelligent persuasive interfaces, instead, the overall goal is to produce an effect on humans, to influence their beliefs, their attitudes and eventually their actions and overall behavior. The area of intelligent persuasion has the potential to change the picture radically in the world of advertising and of social influencing. Computer-based systems can be flexible, and starting from goals they have to pursue, they can take into account the situation and the specific target, adapt the messages in appropriate ways and assess the outcome. In addition, the availability of very large amounts of data which can be exploited also in real time provides unprecedented possibilities. It is easy now to predict the following developments for the advertising sector: reduction in time to market and extension of possible occasions for advertisement; overall reduction of off target messages, eliminating the less relevant for the individual in a given situation; more attention to the wearing out of the message and to the need for planning variants and connected messages across time and space; contextual personalization, on the basis of audience profile and dynamic model (emotional state, beliefs, goals, etc.) and situational information; interactivity; audience reaction monitoring and system feedback on message effectiveness. In the Per Te project we have explored several areas concerned with intelligent persuasion. One topic is concerned with getting the attention and evoking a desired concept by means of original linguistic expressions. A main theme is the automatic production of flexible creative messages [2]. The approach is based on developing specific techniques, mostly corpus based, for producing variations of given expressions [3]. Another theme is concerned with ambient intelligence and peripheral displays. A novel technology is able to indirectly but purposefully impact on the behavior of a co-located group of people. Continuous recognition of each individual's focus of attention and activity is used to drive a tabletop interface aimed at facilitating group interaction during a brainstorming-style activity [4}. We believe one of the ultimate criteria for interactive systems' acceptability is going to be the moral implications of their actions. Persuasion is a sensitive topic and we believe ethics has to be studied from now, in view of granting systems will be accepted by users. We have conducted experiments based on the approach of moral dilemmas, so to understand what persuasion means are acceptable in stressed contexts [5]. In conclusion the potential of intelligent persuasion systems is great, not only for commercial applications but also with a societal view: for influencing social attitudes of people and leading to better social behavior.