Artificial Intelligence (AI) is on everyone’s minds. It’s in articles about the future of work and on billboards promising responsible AI for healthcare. It’s in academic debates about the ethics of machine-decision mechanisms, and in Youtubers streams about singularity.
TEXT Sandra Rodriguez, Kreativdirektorin und Produzentin für interaktive Kunstwerke sowie Soziologin für neue Medientechnologie
Whether real or imagined, we are surrounded by promises of an imminent, inevitable future, where artificial intelligence (AI) and machine-learning tools become part of our social fabric. Yet, such vague promises come with inevitable dilemmas – what are the actual opportunities, redflags and dangers of AI? Should we embrace new developments? What do we know of them? How will AI affect our work, leisure, and economy? And if AI is made to sound human and even look human, thanks to virtual avatars quick developments, can we be fooled into believing who is a human? Vague promises are inspired by fiction. But they also influence hopes, fears and false expectations about what AI should and will become. CHOM5KY vs CHOMSKY: a playful conversation on AI, is an interactive and immersive experience that invites visitors to debunk AI’s myths, hopes and promises.
Surely, academics could complain that the general public doesn’t “understand AI well enough”. But that wouldn’t be fair: the concept of AI is still theoretically challenged (with debates on how it should be defined and what it should encompass). AI is also not one, but an umbrella of different technologies, some of which we use everyday: AI is part of our digital assistants, of algorithm-programs used to help predict who gets jobs. It is part of what helps create deep fakes videos of politicians and famous actors. And as Big tech companies suggest AI “needs time to become what it may”, in the meantime, ghost-workers are hired around the world to play the role of chatbots on websites (a phenomenon called pseudo-AI). Others are hired to help label images for computer vision, or tweak mistakes in natural language processing (NLP). In other words, humans imitate bots, so that AI can live up to the expectations other humans create for it.
In all this confusion, artificial intelligence is made to feel mysterious, specialized and out-of-reach. If we don’t understand what AI is, if we don’t understand the promises made, how can we take part in steering its future? Now is the time to demystify AI. Now is the time we all take part in the conversation.
CHOM5KY vs CHOMSKY – A playful conversation on AI
When addressing social issues, artistic work come as a simple way to ground common assumptions. Art can make scientific developments tangible to a wider audience. It can help individuals grasp general concepts, while raising questions on complex issues. CHOM5KY vs CHOMSKY is an immersive, virtual reality and AI led experiences that strives to do just this: to open a conversation on AI, by inviting audiences to interact, play, and test AI systems.
Technological developments of today are linked to our socially constructed imaginaries. By inviting visitors to peek under the hood of AI, we forge paths to reimagine collective AI future. In the case of AI, for instance, a strong imaginary was built around the possibility to meet and interact with a “life-like” entity that can imitate a human presence so well, that one mistakes it for a real person - sometimes, such entities are depicted like robots. Sometimes they are virtual. One can here think of popular science-fiction books and films – hollywood’s movies A.I., Ex-Machina, Her, etc. In real life, artistic projects are also increasingly using virtual tools and AI to give the audience gt a sense of speaking with someone from the past – such as the Salvador Dali AI, that invites museum visitors to meet an artificial Dali to learn about his life, or similar entities such as the AI-recreated Audrey Hepburn, JFK or Nixon.
CHOM5KY vs CHOMSKY lures visitors with a similar promise. What if we could replicate one of today’s most famous minds? What if you could talk to it, interact with it, and question complex notions like the limitations of GAN, NLP and machine learning systems?
Creating a unique guide: CHOM5KY_AI
In my previous work as an artist, I have often mis-used technology, to debunk myths about it - tracking a web-audience to discuss data tracking, using VR to discuss perception of reality. Since the goal of CHOM5KY vs CHOMSKY is to use AI to talk about AI, it meant to create a host that could simultaneously talk about AI, while discussing complex related subjects: intelligence, free-will, creativity or biases in machine-programming.
“In the emerging cognitive sciences, we really are in a kind of pre-Galilean stage. We don’t know what we’re looking for anymore than Galileo did. There’s a lot to learn from that.”
Noam Chomsky im Interview mit Yarden Katz – the Atlantic
Interviewed by Yarden Katz – the Atlantic
Our AI guide, CHOM5KY, was built on the traces of a world-famous intellectual, known for his life-long research on the human mind – professor Noam Chomsky. Chomsky is a public figure. Known for his political statements, hundreds of books and speeches, the famous linguist has been interviewed on every subject over 60 years, approving interview requests from journalists, students, farmers, artists and scientists alike. His talks, speeches, and debates have been recorded, redacted, transcribed and uploaded, making him one of the most digitized living intellectuals. What’s more, Chomsky is also a social critic and a political activist who warns us about the dangers of big tech corporations when they limit our human capacity for creativity. He is additionally perhaps best known for his theory on Natural Language, a theory that partly inspired “Natural Language Processing” – NLP, the automatic manipulation of language and text that allows us to interact with chatbots. In other words, Chomsky’s extensive wake of digital traces creates a pool of data deep enough to build an AI system. But the content of his digital traces also hold important lessons on his legacy: making him a perfect case-study and a perfect guide to help us demystify AI.
Creating systems to debunk systems
The experience CHOM5KY vs CHOMSKY is a co-production between the National Film Board of Canada, and the Berlin based studio SCHNELLE BUNTE BILDER. The NFB, as a public institution, has the mandate to support artwork that debunks subjects relevant to contemporary social lives. On their end, SCHNELLE BUNTE BILDER is a bustling creative studio of designers, artists, developers, and innovators, who have created award-winning museum exhibits, large scale installations and digital communication strategies that question and think of the media uses in the context of the content and the space they are connected with.
Together, we joined forces in this international collaboration, to make sure we could create an experience that was accessible to a diversified and wide audience. We wanted to craft a space where visitors’ creativity and curiosity could be put front and center – as to really help spark thoughts on issues relevant to our collective futures with AI, while ensuring visitors could also find the experience weird, creative and fun enough to wonder about how they may want to imagine the future of interacting with “AI” and Virtual Entities.
The result is a multi-user experience, where visitors use virtual reality headsets to enter an AI world. They are greeted by CHOM5KY, the AI entity who will act as their guide. The experience merges STT, NLP conversational GPT-2 systems and Unreal game engine, to create a lifelike space where visitors can meet, talk and play with an AI, but also are constantly made aware of each other’s presence and connections. As they explore this world, different forms of play are used in increasing measures to helps visitors progress in their understanding of: a) how AI runs on data; b) how the NLP system looks to understand a sentence and match it to a response; c) how changing parameters of a system may affect its output and d) invite more philosophical questions on what we know of human intelligence, and where it may or may not intersect with that of machines.
To be sure, CHOM5KY vs CHOMSKY is not a deepfake of professor Noam Chomsky. It does not aim to make visitors fear or embrace AI. We want people to question, inquire how AI works, and see where they may imagine other ways of applying it to their lives.
A prototype for the experience was showcased during the Sundance New Frontier Festival in 2020.
“The world is a very puzzling place. There are puzzles, everywhere you look. If you‘re not willing to be puzzled, you just become a replica of someone else‘s mind.”
4. – 20. NOV ab 12:00 - 20:00, KINDL – Zentrum für zeitgenössische Kunst, in Person, Deutsch | Englisch
CHOM5KY vs CHOMSKY: A Playful Conversation on AI
SCHNELLE BUNTE BILDER
1. NOV ab 17.00, Max-Delbrück Zentrum für Molekulare Medizin, in Person, Englisch
Vernissage: The Gene Project, Summet Rohilla
Max-Delbrück Zentrum für Molekulare Medizin
2. NOV 19.00, Zeiss-Großplanetarium, in Person, Englisch
Einstein in the Dome: Project KopfKino
Stiftung Planetarium Berlin, Einstein-Stiftung Berlin
5. NOV 9.30, CAMPUS, in Person, Englisch
Artificial Intelligence: Ethical Questions
Cluster of Excellence – Science of Intelligence
4. NOV 17.00, Studio Roman Lipski, in Person, Deutsch | Englisch
Quantum Physics Meets Contemporary Artistic Practice
Humboldt-Universität zu Berlin
7. NOV ab 14.00, AI Campus, in Person, Englisch
AI, Future, EUROPE, merantix
Sandra Rodriguez, Ph.D., is a creative director and producer for interactive, VR, XR, AI and immersive art works, as well as a sociologist of media technology. Her work as director and producer have garnered multiple awards. Her recent work spans from AI dance performance, multi-user XR theater and large scale AI exhibition. She is a lecturer at Massachusetts Institute of Technology since 2017, where she founded HackingXR, MIT’s first official course on immersive media production.