O Nicholas Carr tem um interessante post discutindo uma entrevista de Frank Schirrmacher (editor de ciência e cultura do Frankfurter Allgemeine Zeitung) no The Edge.  Entre as várias respostas à entrevista, surge um texto (da autoria de John Bargh, chefe do Automaticity in Cognition, Motivation and Evaluation Lab de Yale) que me lembra uma discussão interessante que já tivemos aqui:

Schirrmacher is quite right to worry about the consequences of a universally available digitized knowledge base, especially if it concerns predicting what people will do. And most especially if artificial intelligence agents can begin to search and put together the burgeoning data base about what situation (or prime) X will cause a person to do. The discovery of the pervasiveness of situational priming influences for all of the higher mental processes in humans does say something fundamentally new about human nature (for example, how tightly tied and responsive is our functioning to our particular physical and social surroundings). It removes consciousness or free will as the bottleneck that exclusively generates choices and behavioral impulses, replacing it with the physical and social world itself as the source of these impulses. …

It is because priming studies are so relatively easy to perform that this method has opened up research on the prediction and control of human judgment and behavior, ‘democratized’ it, basically, because studies can be done much more quickly and efficiently, and done well even by relatively untrained undergraduate and graduate students. This has indeed produced (and is still producing) an explosion of knowledge of the IF-THEN contingencies of human responses to the physical and social environment. And so I do worry with Schirrmacher on this score, because we [are] so rapidly building a database or atlas of unconscious influences and effects that could well be exploited by ever-faster computing devices, as the knowledge is accumulating at an exponential rate. …

More frightening to me still is Schirrmacher’s postulated intelligent artificial agents who can, as in the Google Books example, search and access this knowledge base so quickly, and then integrate it to be used in real-time applications to manipulate the target individual to think or feel or behave in ways that suit the agent’s (or its owner’s) agenda of purposes.

Mais gente do que você imagina acredita nisso.  Bem, a conclusão do Carr é frightening:

The Web has been called a “database of intentions.” The bigger that database grows, and the more deeply it is mined, the more difficult it may become to discern whether those intentions are our own or ones that have been implanted in us.”