Global workspace
A central AI system area where information is processed and made available for use by other AI processes.
Think of a global workspace like a central bulletin board in an office. When information comes in, like a new task or a piece of data, it gets posted on this board. Then, different departments or workers (other parts of the AI) can look at the board, grab the information they need, and act on it.
For example, if an AI is writing a story and needs to remember a character's name from earlier in the text, that name might be placed in the global workspace. The part of the AI responsible for character consistency can then access it from the workspace to make sure the name is used correctly.
The global workspace theory, originally from cognitive psychology, proposes a central processing area in the brain where information is broadcast to various specialized modules. In AI, this concept often translates to a shared memory or communication hub where intermediate results or salient information are made accessible to different components of a larger system, such as a large language model or a multi-agent system.
This architecture allows for flexible integration of information and coordination between different processing units. Tradeoffs include potential bottlenecks if the workspace becomes overloaded and the complexity of managing access and updates to shared information.
Recent research, like that from Anthropic, is exploring how to understand and potentially influence the internal states of AI models. This involves looking at how information is processed and made available within the AI, which aligns with the concept of a global workspace.
- Anthropic Can Now Read Claude’s MindAnthropic · The AI Daily Brief ·
- Anthropic Claude
- Anthropic's research investigates the internal workings of their Claude models, which may involve understanding how information is shared and processed within the AI's architecture, akin to a global workspace.
- [object Object]
- [object Object]
The term originates from Bernard Baars's Global Workspace Theory in cognitive psychology, which describes how consciousness arises from information broadcast to a wide range of unconscious cognitive processes. Source ↗
Ask an AI chatbot to summarize a long document you provide, then ask it a specific question about a detail from the middle of the document. Observe if it can recall and use that specific piece of information, which relies on its ability to make that detail available in its processing.
The Playbook: reusable how-tos, prompts, configs and tips pulled out of the news.
- PromptRole-play a student struggling with math for AI tutor simulationSimulate a student's math struggles and uncertainty for AI tutor evaluation. · 5 min
- TemplateReusable workflow templates for healthcare teamsTeams can create and share templates for common workflows like documentation or prior authorizations. · 15 min
- TemplateTemplate for building an AI tool tailored to your workA structured approach to defining and building a custom AI tool for specific business needs. · 15 min
- TemplateEvals CSV prompt list for agent skillsA CSV file template for listing prompts used in evaluating agent skills. · 20 min
- HeuristicDouble coding agent task completion with docs linksProviding coding agents with a documentation link significantly increases their task completion rate. · 2 min
Questions people ask
- Is the global workspace the same as the AI's memory?
- It's related to memory, but more specifically it's a system for making information accessible to different parts of the AI for active processing and decision making.
- Does every AI have a global workspace?
- Not all AI systems are designed with an explicit global workspace. It's a theoretical concept that helps explain how complex cognitive functions can emerge from simpler processing units.
First seen on Subvolts Jul 7, 2026. Updated Sep 16, 2026. Written by an AI model from the items above and reviewed against them weekly; report an error.