Nineteenth International Conference on Technology, Knowledge, and Society

Abstract

The traditional definition of artificial intelligence as “imitation” of human intelligence is simplistic. It does not take into account studies of intelligence in cognitive and behavioural sciences such as Sternberg’s triarchic theory, Garder’s theory of multiple intelligence or least known theories such as Lonergan’s dynamic and integrative epistemology. On the one hand, experiments like the Turing machine reduce intelligence to answering questions while AI software reduces intelligence to mathematical operations or logical inference. This paper argues that mathematical operations and logical inference are based on the fixation of meaning and reference through predicate calculus hence the predilection for fixity, formalization and quantification. Human intelligence involves four integrated but differentiated operation of experiencing (pathos), understanding (logos), judging (ethos) and action (praxis). These operations occur at four levels of consciousness, namely, the empirical, the intellectual, the rational or evaluative and responsible levels and appeal to resources such as attention, intelligence, reasonableness and responsibility. Henceforth, the question Whose Intelligence is the wrong question, a better question would be WHICH intelligence because AI in in its current state limits itself to the second level of consciousness i.e. the intellectual (logos) and ignores the other three dimensions and the implicit dimensions of human knowing. There is a need to question the epistemology that underlies current theories and practices in AI for future improvements.