The I in AI and the meaning in information (IAMI)
Organizer: Terrence Deacon, UC Berkeley [deacon@berkeley.edu]
Following the theme of the 2019 Summit this mini conference will focus on what AI is, isn’t, could be, and shouldn’t be, as well as to consider how it is related to mental processes produced by brains and to the different concepts of information. The question ”Where is the I in AI?” can be interpreted in two ways: Where is the intelligence or where is the agency? The question of intelligence addresses the problem of meaning. Can our most sophisticated AI systems merely be reduced to automated look-up tables of correlations and categorizations of enormous scope and subtlety, or do the embody meaningful knowledge about the world? The question of agency also addresses this problem but additionally begs us to consider the locus of the goals and purposes of these systems. Are they merely parasitic of their designer’s goals? To what extend are self-training machine learning approaches capable of autonomous agency? Is the unmonitored autonomous operation of an AI-based machine process a form of agency, or is this superficial appearance Is there some fundamental difference between AI and neurologically based intelligence that makes true autonomous agency (and subjective experience) unattainable for machine-based systems?
The second part of this year’s theme statement is addressed in the introduction to the President’s statement:
“The information revolution—initiated in the middle of the 20th Century by the theoretical insights of Claude Shannon, Alan Turing, John Von Neumann, and innumerable others—was made possible by setting aside issues of meaning and significance in order to derive precise theoretical tools for measuring and controlling information. But with the maturity of information technologies, these once ignorable attributes of information have now become matters of global importance. We are rapidly racing toward a future created in the image of this mindless sense of information. … Our understanding of how meanings and values emerge from the web of information-sharing, data analysis, and automated communications remains under-developed and often … considered superfluous.” [excerpt from the President’s statement]
Relevant topics for this mini conference include:
• Formalizing the relationship between syntactic and semantic concepts of information.
• Identifying similarities and critical differences between machine-based and neurological information processing.
• Using insights from brain organization and function to design machine learning, perception and action.
• Alternative measures of complexity that are relevant for living and mental processes
• The transition from non-normative physics and chemistry to information about the relevant environment for a living organism.
• Origins and nature of biological information at the cellular and molecular level.
• The emergence of semiotic properties at the origin of life.
• Cognitive cyborg technology: Integration of AI-based prostheses with brains
Title, author information, and abstract should be written in English and should be no longer than 1 page.
For questions concerning this mini conference contact deacon@berkeley.edu