Overview
In a paper submitted on 11 September 2026 (arXiv:2609.30291, cs.AI), Joseph Sifakis argues that autonomous systems represent the ultimate stage in the development of artificial intelligence. The work explains the technical challenges that demand a combination of connectionist AI and symbolic AI, and makes the case for integrating AI with systems engineering.
A Framework for Autonomous Systems
The paper presents a comprehensive framework for designing and evaluating autonomous systems. It is built on a generic agent architecture that characterizes behavior as the composition of cognitive functions organized around a long-term memory containing the agent's evolving knowledge.
Key Technical Challenges
Sifakis addresses the implementation of the architecture's fundamental features, including:
- Linking sensory data to structured data stored in memory
- Decision-making related to achieving the agent's goals and planning for them
- Coordination among agents to combine individual and collective intelligence
Trustworthiness Beyond Behavior
A central argument of the paper is that agent trustworthiness differs from that of traditional systems. It is not limited to behavioral properties but includes an essential cognitive dimension: the validity of an agent's decisions depends on how it uses its knowledge. The paper outlines avenues for developing methods to evaluate agent trustworthiness along these lines.
Critical Assessment
The work concludes with a critical assessment of the substantial gap between the aspirational vision of autonomous multi-agent systems and the current state of the art.
Publication Details
- Author: Joseph Sifakis
- Submitted: 11 September 2026
- Subject: Artificial Intelligence (cs.AI)
- Cite as: arXiv:2609.30291 [cs.AI]
- DOI: https://doi.org/10.48550/arXiv.2609.30291
via ArXiv AI