[Paper Review] From Control to Foresight: Simulation as a New Paradigm for Human-Agent Collaboration

Gaole He, B. Lim|arXiv (Cornell University)|Mar 12, 2026

Human-Automation Interaction and Safety0 citations

TL;DR

The paper advocates simulation-in-the-loop to provide foreseen future trajectories in human-LLM agent collaboration, transforming intervention from reactive to proactive exploration.

ABSTRACT

Large Language Models (LLMs) are increasingly used to power autonomous agents for complex, multi-step tasks. However, human-agent interaction remains pointwise and reactive: users approve or correct individual actions to mitigate immediate risks, without visibility into subsequent consequences. This forces users to mentally simulate long-term effects, a cognitively demanding and often inaccurate process. Users have control over individual steps but lack the foresight to make informed decisions. We argue that effective collaboration requires foresight, not just control. We propose simulation-in-the-loop, an interaction paradigm that enables users and agents to explore simulated future trajectories before committing to decisions. Simulation transforms intervention from reactive guesswork into informed exploration, while helping users discover latent constraints and preferences along the way. This perspective paper characterizes the limitations of current paradigms, introduces a conceptual framework for simulation-based collaboration, and illustrates its potential through concrete human-agent collaboration scenarios.

Motivation & Objective

• Identify limitations of current pointwise human–agent interaction paradigms.
• Define a conceptual framework and design space for simulation-based collaboration.
• Demonstrate how simulated futures support foresight, not just planning, in multi-step tasks.
• Highlight practical implications and trade-offs for enabling serendipity and latent constraint discovery.

Proposed method

• Define four core concepts: agentic workflow, action space, simulation, and simulated impact.
• Describe how simulation externalizes internal future exploration to aid sensemaking.
• Provide an illustrative multi-city trip planning scenario with multiple futures and annotated outcomes.
• Outline a design space with lookahead depth, exploration breadth, and granularity considerations.

Figure 1. Comparison of interaction paradigms at a decision point.

Experimental results

Research questions

• RQ1 How do current pointwise interaction paradigms limit human–agent collaboration in long-horizon tasks?
• RQ2 What is the design space for simulation-based interaction, and how do trade-offs affect reasoning and trust?
• RQ3 Can simulation-in-the-loop enable proactive human–agent exploration and discovery of latent constraints and needs?

Key findings

• Simulation-in-the-loop transforms intervention from reactive guesswork to proactive exploration.
• Users can compare multiple simulated futures with annotated outcomes to make informed trade-offs (e.g., time vs. cost, delay risk).
• Simulation reveals latent constraints and serendipitous options that would be missed under traditional paradigms.
• Design choices (lookahead depth, breadth, and granularity) critically shape user cognition, trust, and cognitive load.