Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Interacting with Thoughtful AI

Xingyu Bruce Liu, Haijun Xia, Xiang Anthony Chen

TL;DR

Thoughtful AI proposes a continuous-thinking AI paradigm that moves beyond turn-based, input-output interactions to support proactive, evolving thought processes and a shared cognitive workspace with users. Four core traits—intermediate medium, full-duplex reasoning, intrinsic driver, and shared cognitive space—enable AI to generate, develop, and selectively communicate its thoughts, enabling continuous cognitive alignment and more dynamic collaboration. The authors illustrate the concept with two projects, Inner Thoughts and ThinkaloudLM, showing improvements in interaction quality and transparency. They discuss implications for HCI, potential designs, and a research agenda addressing traits, downsides, architectures, and the broader impact on human thinking and interaction paradigms.

Abstract

We envision the concept of Thoughtful AI, a new human-AI interaction paradigm in which the AI behaves as a continuously thinking entity. Unlike conventional AI systems that operate on a turn-based, input-output model, Thoughtful AI autonomously generates, develops, and communicates its evolving thought process throughout an interaction. In this position paper, we argue that this thoughtfulness unlocks new possibilities for human-AI interaction by enabling proactive AI behavior, facilitating continuous cognitive alignment with users, and fostering more dynamic interaction experiences. We outline the conceptual foundations of Thoughtful AI, illustrate its potential through example projects, and envision how this paradigm can transform human-AI interaction in the future.

Interacting with Thoughtful AI

TL;DR

Thoughtful AI proposes a continuous-thinking AI paradigm that moves beyond turn-based, input-output interactions to support proactive, evolving thought processes and a shared cognitive workspace with users. Four core traits—intermediate medium, full-duplex reasoning, intrinsic driver, and shared cognitive space—enable AI to generate, develop, and selectively communicate its thoughts, enabling continuous cognitive alignment and more dynamic collaboration. The authors illustrate the concept with two projects, Inner Thoughts and ThinkaloudLM, showing improvements in interaction quality and transparency. They discuss implications for HCI, potential designs, and a research agenda addressing traits, downsides, architectures, and the broader impact on human thinking and interaction paradigms.

Abstract

We envision the concept of Thoughtful AI, a new human-AI interaction paradigm in which the AI behaves as a continuously thinking entity. Unlike conventional AI systems that operate on a turn-based, input-output model, Thoughtful AI autonomously generates, develops, and communicates its evolving thought process throughout an interaction. In this position paper, we argue that this thoughtfulness unlocks new possibilities for human-AI interaction by enabling proactive AI behavior, facilitating continuous cognitive alignment with users, and fostering more dynamic interaction experiences. We outline the conceptual foundations of Thoughtful AI, illustrate its potential through example projects, and envision how this paradigm can transform human-AI interaction in the future.

Paper Structure

This paper contains 23 sections, 3 figures, 1 table.

Table of Contents

  1. Introduction
  2. What makes AI "Thoughtful"?
  3. Traits of Thoughtful AI
  4. Intermediate Medium
  5. Full-duplex Process
  6. Intrinsic Driver
  7. Shared Cognitive Space
  8. Implications for HCI
  9. From Passive Respondents to Proactive Participants
  10. Continuous Cognitive Alignment
  11. Beyond Turn-based Interaction
  12. Messy, Fragmented and Informal Interaction
  13. Example Projects
  14. Inner Thoughts
  15. ThinkaloudLM
  16. ...and 8 more sections

Figures (3)

  • Figure 1: Rethinking human-AI interaction paradigms: (A) Traditional AI is reactive, responding only when prompted. (B) Thoughtful AI thinks continuously, proactively generating, iterating, and allowing users to interact with its thoughts.
  • Figure 2: Conversational Agents with Inner Thoughts: AI generates a train of thoughts and evaluates them based on their intrinsic motivation to participate.
  • Figure 3: ThinkaloudLM: AI generates intermediate, fragmented thoughts in parellel to user input.

Theorems & Definitions (1)

  • definition 1