Hi there!

I’m Romy - a designer, artist and technologist, interested in experimenting with new media and technologies to create interactions and immersive experiences that bridge the space between the physical and the digital.

Find out more:
︎ Information
︎ Linkedin
︎ Resume

Featured            UX Design            Code            WorldBuilding


MIT Media Lab | 2020
Ethan Zuckerman
Individual work Duration: 1.5 month Tools Used: Adobe XD, HTML

Background & initial speculative ideas

For the class “Fixing Social Media”, I was interested in addressing the topic of political polarization online. I wanted to create a way for people to actively try and learn from the other, and proactively extract themselves from echo chambers.

How can we extract ourselves from the echo chambers we’ve been so used to,
and expose ourselves to people from different backgrounds/beliefs?

Why are datasets not designed to allow us to learn about another stranger?

How can data about liked/shared pics of food become a connector?

How can data from twitter activity become a connector?


Political debates are tense and difficult to handle - sometimes we can see participants becoming stubborn, and unwilling to change their opinion. This narrow-mindedness in discussions reinforces political polirization. This project aims to create a safe space for lebanese people to have a debate on controversial opinions, and get to a productivedialogue.

How might we design a healthy online space to debate controversial political topics?



Problem space

User interviews
Phone debates
Journey mapping

Design solutions
Prototypes (2)
User flow


User evaluation


Phase 1: Survey (53 respondents)

To  understand user sentiment towards debates, a survey was designed with several behavioral questions with support from online sources. With 53 respondents, these were the 3 most insightful responses:

What makes a bad debater?

What would you improve in your debating skills?

What is most important in a moderator?

Phase 2: Interviews (4 participants)

I continued to interview people that had used Lebanocracia.org. Since that platform did not allow for discussions, I was hoping to get their personal sentiment towards political debates in Lebanon. I asked the following questions:

- Do you like debates? Why & why not?
- What about political debates in Lebanon? Why & why not?
- What do you think is the crucial problem that prevents productive debates?
- Why would it be beneficial to even have debates?

Maya, 27
Architect, Tripoli,
Sunni Muslim

“I get too angry, and give up as soon as I see it becoming personal”

Michel, 65
Dirctor, Mount Lebanon, Maronite

“I would personally love to be able to have a productive conversation with an opponent and spread it to more people”

Zeina, 32
Designer, Baalbeck,

“The thing that I expect everytime in Lebanon, is pure stubborness because everyone still relies their authority’s ideals”

Rami, 35
Engineer, South Lebanon, Shiia

“It’s too bad sometimes because debates actually allow people to learn about important issues/subjects”

Phase 3: Debates moderated by myself (3x2 participants)

To have a more direct look at how debates are conducted, what is are explicit problems and implicit, I decided to moderate 3 virtual debates with the following topics:

1. Investment in tourism or local production?
2. Government technocrat or religiously tied?
3. House by the beach or in the mountains?

The problems observed were the following: Impulsive responses, No facts, Patronizing, Mansplaining, Making the argument personal, Misinformation, Jumping to conclusions based on someone’s original argument.

Video debates:


Seeing someone’s face gives room for more compassion.

More sense of accountability when you know the opponent is a “person”


There is a bit more restraint in argumentation (if the participants didn’t know each other)

People speak over each other, and the job of the moderator is more difficult.

Main Research Insights


People believe the number one cause for a bad debate is baseless arguments and misinformation.


A moderator’s role should be less about rephrasing positions, and more about controlling the direction & pace.


Texting instead of video debates has less of a sense of accountability, but could allow more freedom of speech.


People don’t generally admit that they are bad listeners.


Because of the ego, moderators should be the ones to point out tricky subjects such as World View Attachment.


From interviews, people showed enthusiasm towards a healthy debate space because it could be educational, and could be an example to become a better debater.


Design Iterations

To solve for the 6 insights found above, I seperated them into 3 buckers: Solving for Misinformation, Solving for impulsivity, Promoting Openmindedness. I then got some feedback on each proposal from MIT students and people from Lebanon.

Solving for Misinformation

Through exposure to facts before or during the debate.

Solving for Impulsivity

Through rating that is either built in the text chat, or as a visualization through the debate.

Open mindedness

Through an AI moderator that would notice World View Attachment, and preventing prejudices by exposing participants to the belief of their demographics.

“helpful and education - but information might be lost afterwards”

“is there a way to link to a source during the debate”

“although you immediately see the sentiment of your opponent on your argument, there is no accountability throughout the debate if it disappears with the debate”

“reading the guidelines to a good debate is good - don’t you think the AI moderator will be distracting during the debate?”

“distracting but could ground the conversation in factual information”

“is the selection of the data viewable by the other party?”

“having this direct sense of accountability  allows people to view their contribution, and promotes openness and reflection”
“it’s like a summary landscape of arguments”

“love the ability to compare yourself to your own demographics”

User Flow

The 4 main sections of the platform are: logging in and filling information, customizing debate topic/question/opponent & scheduling, the debate portion (30 minutes in duration), and sharing the debate to social media platforms.

Design Features


Understanding your demographic:
to avoid prejudices,
and a set the scene for the conversation.


A central buffer zone with left & right for participants:
intended for reflection, and awareness of physical presence of other.


AI moderator: that collects articles/datasets.


Rating system: to enforce reflection
& openmindedness with every argument


Sharing: to impact more people, participants are able
to share their debate chat history as an animated
visualization on social media platforms.


Walk through dataDialogue.


Sharing dialogue history on social media platforms to target specific groups of people and reach to a broader audience.