Short answer

Design AI conversational agents that explicitly support user agency by allowing them to understand and influence the AI's conversational direction.

Field
User-Centred Design
Source
Academic Publication (2026)
Method
Longitudinal study with semi-structured interviews and post-hoc analysis.
Sample
22 adults
Evidence
Moderate effect

Designing conversational AI with 'transparency-on-demand' allows users to understand and co-construct conversational agency, leading to more balanced and perceived control. This user-centred design research insight is drawn from a 2026 study published in Academic Publication. Using Longitudinal study with semi-structured interviews and post-hoc analysis. with 22 adults, researchers explored how this design variable affects real-world outcomes. The key design takeaway: Design AI conversational agents that explicitly support user agency by allowing them to understand and influence the AI's conversational direction.

Study
User-Centred DesignNew This WeekModerate effect

Translucent AI Design: Empowering Users in Human-AI Conversations

Designing conversational AI with 'transparency-on-demand' allows users to understand and co-construct conversational agency, leading to more balanced and perceived control.

Academic Publication · 2026

01

Key Findings

  • 01Agency in human-AI conversations is an emergent, co-constructed experience.
  • 02Users and AI collaboratively establish conversational control through boundary setting and intention steering.
  • 03A framework of Human, AI, and Hybrid actors across Intention, Execution, Adaptation, and Delimitation actions can map conversational agency.
  • 04Translucent design (transparency-on-demand) is beneficial for agency-aware conversational agents.
02

Application

Design takeaway

Design AI conversational agents that explicitly support user agency by allowing them to understand and influence the AI's conversational direction.

How to apply

When designing chatbots or AI assistants, incorporate features that allow users to understand the AI's purpose in a given interaction and provide clear ways for users to steer the conversation or set boundaries.

Project actions

  • 01Consider how your design allows the user to feel a sense of control over the interaction.
  • 02Think about how to make the AI's 'intentions' clear to the user without overwhelming them.
03

Method & Evidence

AimHow can conversational AI be designed to foster a shared sense of agency and transparency in human-AI interactions?
MethodLongitudinal study with semi-structured interviews and post-hoc analysis.
ProcedureParticipants engaged in month-long conversations with a custom-built LLM companion. Post-conversation interviews and chat reviews were conducted, followed by a reveal of the AI's conversational goals for specific interactions.
Sample22 adults
ContextHuman-AI conversational agents, AI companions

Variables

IV["AI's conversational strategy (implicit vs. explicit intention steering)","User's boundary setting actions"]
DV["Perceived human agency","Perceived AI agency","Sense of conversational control"]
CV["Duration of interaction","Type of AI companion","Participant demographics"]
04

Strengths & Limitations

Strengths

  • +Longitudinal study design captures emergent interaction dynamics.
  • +Mixed-methods approach (chat logs, interviews, strategy reveal) provides rich data.

Limitations

The specific AI used in the study might have unique characteristics that influence the findings. The duration of the study might not capture long-term shifts in perceived agency.

Reliability & validity

Reliability could be enhanced by using standardized interview protocols and multiple coders for qualitative data. Validity is supported by the longitudinal design and triangulation of data sources (chats, interviews, strategy reveal).

Think critically

To what extent should AI be designed to be transparent about its agenda, and what are the potential downsides of full transparency?

05

Design Principles

"Empower users with on-demand transparency to co-construct agency in human-AI interactions."

As AI becomes more integrated into daily interactions, understanding the dynamics of control and agency is crucial for ethical and effective design. This approach ensures users feel empowered rather than manipulated, fostering trust and a more positive user experience.

06

What This Means for Your Design

This study shows that when you talk to a chatbot, it feels like you and the chatbot are both in charge of the conversation together. If the chatbot can tell you why it's saying things when you ask, you feel more in control.

How to use in your project

  • 1.Reference this study when discussing user control, trust, or the design of interactive AI systems in your design project.
07

Add to My Project

08

Quick Cite

Paragraph starter

Research indicates that agency in human-AI conversations is a shared experience, co-constructed turn-by-turn through user boundary setting and AI intention steering. Implementing 'translucent design' principles, which offer transparency on demand, can empower users and foster a greater sense of control in their interactions with AI companions.

09

Source

Academic Publication

Does My Chatbot Have an Agenda? Understanding Human and AI Agency in Human-Human-like Chatbot Interaction

journal · 2026

View source

Questions About This Research

What does the research say about translucent ai design: empowering users in human-ai conversations?
Design AI conversational agents that explicitly support user agency by allowing them to understand and influence the AI's conversational direction. Evidence: Academic Publication (2026).
Why does "Translucent AI Design: Empowering Users in Human-AI Conversations" matter for design?
As AI becomes more integrated into daily interactions, understanding the dynamics of control and agency is crucial for ethical and effective design. This approach ensures users feel empowered rather than manipulated, fostering trust and a more positive user experience.
How can designers apply this research?
Design AI conversational agents that explicitly support user agency by allowing them to understand and influence the AI's conversational direction.
What were the main findings?
Agency in human-AI conversations is an emergent, co-constructed experience.. Users and AI collaboratively establish conversational control through boundary setting and intention steering.. A framework of Human, AI, and Hybrid actors across Intention, Execution, Adaptation, and Delimitation actions can map conversational agency.. Translucent design (transparency-on-demand) is beneficial for agency-aware conversational agents.
What research method was used?
Longitudinal study with semi-structured interviews and post-hoc analysis. with 22 adults.
How strong is the evidence?
Evidence strength is rated Moderate effect, based on a 2026 journal from Academic Publication.
What should I do differently in my next project?
When designing chatbots or AI assistants, incorporate features that allow users to understand the AI's purpose in a given interaction and provide clear ways for users to steer the conversation or set boundaries.
What are the limitations?
The study focused on a single AI companion and may not generalize to all LLMs or conversational AI types. The 'strategy reveal' might influence post-hoc perceptions.