How to Define Intelligence and Consciousness for In Silico and Organoid-based Systems?
A global initiative is underway to create consensus definitions for diverse intelligent systems, including AI, LLMs, and biological intelligences. Led by Cortical Labs, the biological computing startup known for "Dishbrain," this project brings together scientists, ethicists, and researchers from the UK, Canada, USA, EU, Australia, and Singapore. Their goal is to standardize the language in the rapidly advancing and often contentious field of generally intelligent systems.
The Need for Consensus
The field of intelligent systems encompasses a wide range of technologies and disciplines, each with its own terminology. Fifteen years ago, at least 71 distinct definitions of “intelligence” were identified. Today, the rapid growth of this field makes it impractical for researchers to redefine every ambiguous or imprecise term in each publication. A unified language is essential for the recognition, prediction, manipulation, and construction of cognitive (or pseudo-cognitive) systems, particularly those with unconventional embodiments that differ significantly from natural species.
Previous attempts to establish nomenclature guidelines have been highly specific to certain fields and developed by select experts, limiting broader community engagement. The current effort seeks to involve a wider range of stakeholders to develop a theory-agnostic standard that can be applied across multiple disciplines.
The Approach
Figure 1. Initial key terms, most applicable fields, and core approach toward consensus
The collaboration will define key terms applicable to various fields, such as artificial intelligence, autonomous systems, consciousness research, machine learning, organoid intelligence, and robotics (Figure 1A and 1B). The initiative will use a mixed-method approach with a modified Delphi method, which involves several iterative rounds of consultation and refinement:
- Initial Round: Pre-selected open-ended questions are posed.
- Strategic Refinement and Categorization: Responses are analyzed and categorized.
- Collaborative Consultation: Participants engage in iterative discussions to refine terms.
- Achieving Consensus: If consensus is not reached, a weighted majority voting system will be used to finalize definitions.
This approach aims to create a critical field guide for researchers developing diverse intelligent systems, addressing the current lack of standardized terminology.
Collaboration and Participation
Cortical Labs' Chief Scientific Officer, Brett Kagan, emphasizes the importance of this collaborative effort:
“Ultimately, the purpose of this collaboration is to create a critical field guide for researchers across a broad range of fields who are engaged in the development of diverse generally intelligent systems. In what is a rapidly evolving space, such a guide doesn’t yet exist.”
Professor Ge Wang from RPI, USA, echoes this sentiment, highlighting the limitations of current approaches to multimodal multitask foundation models.
Researchers and scientists in related fields are invited to collaborate and contribute to this nomenclature effort, providing nuanced and consistent language for emerging technologies. Interested collaborators can register at Cortical Labs Nomenclature.
Topics: AI & Digital
