The article proposes to use a generally functionalist approach to create the basis of the conceptual model of Artificial General Intelligence (AGI). AGI is a field of scientific-theoretical and engineering-technological research, focused on the construction and use of computer simulations, models and reproductions of cognitive phenomena of the widest range of life, mental, personal and social manifestations. The methodological basis of AGI is general computer functionalism as a contemporary interpretation of the psychophysiological theory of machine functionalism of H. Putnam of the 1960s. There are collective, conceptual and observant approaches to the formation of general functionalism. A collective approach is the collection, identification, coordination, formalization, systematization, unification, codification of all kinds of functionalist theories. A conceptual approach is the analysis and identification of the main functionalistic characteristics, relationships, patterns, causalities that are invariant with respect to the content of cognitive phenomena. The observation approach allows to evaluate from the position of a person or a social community, immersed in the communicative “waves” of the virtual and real world, the different statuses of technological realizations of general functionalism: ontological, epistemological, logical, linguistic, axiological, aesthetic, ethical and praxeological features of projects of artificial life, artificial consciousness (brain), artificial personality, artificial society. As a basic personal phenomenon, which is defined by a system of functional relations, we take not the pain phenomenon familiar to functionalism but the more productive phenomenon of need. It has an advantage due to the breadth of the scale of phenomenological evidence, epistemological adequacy, ontological foundation. Based on functionalist modeling of need, the so-called “artificial need” arises, which can form the basis for the development of AGI. The author of the article offers one of the options for formalizing needs within the framework of an artificial system based on the general functionalist principle of formalizing cognitive phenomena. The principles of formalization proposed by D. Levin and T.W. Polger, who developed the approach of H. Putnam, are used. The author concludes that the application of this methodology leads us to understanding of an observer in the system of functional relations and considers the AGI as a complexity system.