Role

User Research, Workshop Planning & Analysis

‍

Duration

4 months (Jan.2023 - Apr.2023)

‍

Team

5 Researchers

An Overview 👀

‍

‍

‍

What is Dream Distortion? 🔮

‍

The term "dream distortion," as described by Freud, refers to the process of disguise or alteration that occurs when unconscious material is transformed before entering consciousness in the form of a manifest dream. In simpler terms, it’s when the content of a dream is altered or distorted compared to the original thoughts or emotions that triggered the dream.

Research Question 🔦

‍

‍

‍

To explore the relationship between humans and AI technologies, we posed two core research questions:

1. How does the interpretation of dreams differ between humans and technology?
2. Can AI enhance, improve, or change our human perception and interpretation of dreams?

Participatory Workshop ✂️

‍

‍

‍

Our workshop was structured around 4 sessions with 3 participants in each session. Each participant brought a dream they had, which was interpreted both by humans and by AI, followed by group discussions. For AI interpretations, we used an AI text and image generator, inputting the dream descriptions directly into the tool.

‍

Workshop Materials:

‍

• AI interpretations: Text and image interpretations generated by AI based on the dream descriptions.
• Participant contributions: Discussions, sketches, and mood boards capturing their personal interpretations of their dreams.

‍

Final Website 💻

‍

‍

‍

For our final deliverable, we created a visual website with four main sections:

1. Introduction: Overview of the workshop’s focus, goals, and structure.
2. Key Findings: Insights gained from the workshop sessions.
3. Dream Graphics: Visual representations of the 12 dreams shared by participants.
4. Archive: A PDF containing all the materials we collected, including participant notes, sketches, and AI-generated interpretations.

‍

Click to view full website! 🔍

Website Breakdown 🔨

‍

‍

‍

Introduction

‍

This section introduces the purpose of the workshop, discussing how we structured the sessions to compare human vs. AI interpretations of dreams.

‍

Key Findings

‍

We uncovered several interesting insights:

• Humans interpret dreams differently: The emotional and subjective nature of human interpretation contrasted with the literal, data-driven nature of AI.
• People have higher expectations from AI interpretation: Participants were often surprised by how simplistic or straightforward AI-generated interpretations were, given AI’s perceived intelligence.
• AI overthinks: The AI would sometimes create overly complex or far-fetched explanations for dreams, reflecting its attempt to synthesize patterns and information.
• Multiple layers of distortion: Dream distortion exists on many levels—first in the dream itself, then through human and AI interpretation.
• Stereotypes in AI: AI interpretations often revealed biases or stereotypes learned from data, providing a distorted or generalized view of dream content.

‍

Dream Graphics

‍

We visually represented 12 dreams, blending elements from participants’ descriptions, sketches, and AI-generated images. Each graphic showcased the dual nature of dream interpretation by combining human and AI visual outputs.

‍

Archive

‍

The archive contains raw materials from the workshop:

• Original dreams submitted by participants.
• Notes from participants and notetakers.
• Participant sketches and mood boards.
• AI text and image interpretations.