More selected projects

 

Wander [001] - The future nomad

Wander [001] is an AI-based multi-media public art project that includes the interactive AI chatbot, CG and an online map.

On Discord or WeChat, participants can send messages to Wander of any real-world locations, and Wander would visit and send sci-fi travel notes and street scenes from a random future time.

Join the discord channel to contact Wander: https://discord.gg/3njPKQJ4V4

produced by: Yuqian Sun

Introduction

Wander [001] is an AI-based multi-media public art project that includes the interactive AI chatbot, CG and an online map. On Discord or WeChat, participants can send messages to Wander of any real-world locations, and Wander would visit and send sci-fi travel notes and street scenes from a random future time. She can also explore the place according to participants’ following action commands. All contents are generated through AI based on real-world location information.

Through public participation, the history of the journey would keep updating the future map. The map would become a record of a huge collaborative Sci-fi fiction collection between humans and AI.

Based on the travel notes that Wander send, I create series of concept art, and mint them into NFT. The incoming of NFT will support the development of this project.

 

  • gallery-image
  • gallery-image
  • gallery-image
  • gallery-image
  • gallery-image
Concept and background research

My projects are developed through discussion about Intelligent Agent (IA) and enclosed contexts: since the data used by AI algorithms are always generated by a human, what is produced is just a remix of human data? Faster, is creativity and aesthetics still in a closed system? What does it mean when human and nonhumans are ‘assemblages or ‘net- works’?Is creativity really an outcome of endless assemblages of different things?

There are conflicting points about if AI increases the aesthetic diversity in culture. According to Lev Manovich’s research(AI aesthetics), algorithm increase gradual automation (semi or full) aesthetic decisions, which means recommendation engines suggesting what we should view. In general, people tend to get more exposed to content that they haven’t think about, so in terms of choices, recommendation system ( and other AI algorithm)helps to increase diversity. However, Lucy Suchman(2006. 239) draw attention to social robots when she states “the fear is less than robotic visions will be realized than that the discourses and imaginaries that inspire them will retrench received conceptions both of humanness and of desirable robot potentialities, rather than challenge and hold open the space of possibilities”. Measuring imaginations and diversities is not the central point to discuss here since it needs quantitative measurement, but the two points are worth researching. Lev’s point focuses more on personal choices, while Lucy’s statement is more general.

The research project grew through my practice of expanding boundaries in various semantic contexts. I call it IA Art: AI is not only used for production (e.g. as a silent painter to produce paintings), but as an actor to reflect on the role of the machine agency and human beings.

I hope this work can provide a crosspoint between reality and virtuality, and turn familiar surroundings into fantasy: all content is generated based on real-world location's information. Each trip would be like a text-based adventure for participants, but there's no fixed choice: empowered by AI technology, players can visit anywhere on earth, and ask the bot to do anything they imagined.

The future map website would document all the travel logs, and sort them into a timeline. This made Wander [001] into a fully open-world interactive fiction collection that created by humans and AI.  

  • gallery-image
  • gallery-image
  • gallery-image
  • gallery-image
  • gallery-image
  • gallery-image
  • gallery-image
  • gallery-image
  • gallery-image
Technical

Wander accpet 2 commands:
Visit: any location on earth
Action: any action
To ensure fluency, each trip allows 5 actions. This can be any action without any limitation, as long as making sense, like invade the plane, search for lives in the ruins, pat a kitten, etc.
Chinese and English versions share a similar workflow.


User sends a message about the location: visit: London
=》Call Baidu/google map searching API, get the GPS location of Xishuangbanna
=》Call Baidu/google map image API, get image from the location
=》Use style transfer API to generate a futuristic images.
=》Use Dreamly.ai continuation writing API to generate travel notes.
=》Send a message back to the user
=》Wait for action message and generate next message
=》Save the log in backend
We did lots of prompt engineering work to make sure Wander actually know who she is and what world she lives in. The way to do that is to put the following text in the prompt:

Science Fiction: Future Nomad's journey on earth
It's not easy to be a nomad in the cyber generation, even as a robot. My name is Wander, I'm not a human being, but a robot girl, slowly travelling the earth. Now there are not many people on the Earth, some land was destroyed by machinery or meteorites, and other is under free development with the progress of science and technology.
Here is [location] in [year]. [variation] I activate the environment detection module and slowly descent my aircraft.


[location] is the location sent by participants. [year] is the random visit year, and [variation] is the environment description that corresponds to the 3 style transfer type:
style 1: I’m surrounded by ruins. According to my electronic detector, there was no one here, and the air was filled with dust and metal particles. 
style 2: My electronic detector shows that the landscape becomes colourful under the influence of radiation.
style 3:The green scenery pass through my electronic camera. The plant life here is extremely luxuriant: foliage similar to the alien life closely surrounds the building, forming the strange scenery on earth.
With the location’s name and description variation, Wander will produce a rich and surprising responses that corresponds to the transferred image.

 

  • gallery-image
  • gallery-image
Future development

During the exhibition, the online map is actually using static data from former travel history. I will finish the online real-time version soon, and make it into the main site of Wander.

Self evaluation

I’m glad to find it attracted over 2000 participants to join the journey. I aim to create a new methodology of human-AI interaction, which provides a narrative context for AI's creativity, then invites humans to immerse in the fictional world with the familiar messaging software.
They create imaginary future earth together, during which, we can meet surprises. One of the examples is a Chinese female visitor who said her boyfriend is in Hongkong, and his nickname is White Dragon(白龙). When she travels to Hongkong with Wander, after sending ' dive into water area ' action,  Wander really met a white dragon among the ruins. This kinds of unexpected accident break the wall between fiction and reality, and it's the result of collaboration between human and AI.
I’m so moved to receive positive feedback from participants. They feel surprised about how their hometown look like in the future (they can recognize the building and even roads in the picture). But there’s still some pity. I find it’s difficult to attract viewers to join the discord channel during the exhibition. During the former exhibition in China, since the Chinese version is based on WeChat (the most common use message app in China), it’s very easy to invite participants to add the bot and start the journey. Discord is not that common, but I appreciate its public channel so people can see logs from former participants, so I didn’t choose other messaging apps like Messenger or WhatsApp. I will definitely send promotional messages on social media, but I will also improve the way of setting up to clarify the platform is used.

References

Global map reference https://2050.earth/
Wechaty tutorial https://aistudio.baidu.com/aistudio/projectdetail/1836012
WeChaty example https://github.com/Lovely-Pig/paddlehub-wechaty
Baidu map api http://lbsyun.baidu.com/index.php?title=viewstatic https://zhuanlan.zhihu.com/p/79627772
Baidu map location search http://lbsyun.baidu.com/index.php?title=webapi/guide/webservice-placeapi

Exhibition
2021 Asia Digital Art Exhibition in Times Museum, Beijing, China
2021 Goldsmiths Computational Arts MA/MFA Degree Show
2021 Chengdu Art Biennale (Coming up in October)