A philosopher and scientist reflects on seven decades of work on the hard problem of consciousness—how mental phenomena relate to brain processes. He describes developing an information-based framework in the early 1960s, defending it against ideological opposition, especially from followers of E.V. Ilyenkov. Despite accusations of revisionism, he published many books and articles on decoding neural correlates of subjective reality. After a 2019 English-language article in a leading neuroscience journal, he received international recognition, including keynote invitations and editorial board memberships. The article argues that neuroscience urgently needs theoretical frameworks like the information paradigm to guide empirical research on consciousness, mental processes, and biological functions.
Creating artificial general intelligence (AGI) that approaches natural intelligence requires integrating phenomenological studies of subjective reality with neuroscientific findings on consciousness. The article identifies key neuroscientific studies of consciousness that could inform the modeling of specific cognitive architectures for AGI. To justify using neuroscientific descriptions of cognitive operations in computer modeling, the author proposes an informational approach that explains the connection between subjective reality and brain processes. Closer collaboration among specialists in methodology, epistemology, artificial intelligence, and neuroscience is a prerequisite for progress.
Creating Artificial General Intelligence (AGI) requires solving fundamental methodological issues that demand input from philosophers specializing in epistemology and philosophy of science. The paper distinguishes between artificial and natural, strong and weak, and general and narrow intelligence. Theoretical difficulties arise in clearly defining general intelligence's properties and its practical implementation. Studies of consciousness are crucial for AGI development, particularly using phenomenological analysis of subjective reality, its value-semantic and operational structures. These issues are directly relevant to constructing new cognitive architectures that go beyond narrow AI, enabling high autonomy and independent problem-solving. Turing's operationalist methodology is limited because it excludes results from consciousness studies; a post-Turing methodology opens significant opportunities for AGI.