A prologue to a timely conversation on AI’s biggest promises and pitfalls. Lured by the possibilities of an emulation of one of today’s most famous minds, we meet and engage with CHOMSKY_AI, an entity under construction. Renowned thinker Noam Chomsky has devoted his career to studying the mind. Drawing from the arsenal of digital traces he’s left behind, CHOMSKY_AI challenges us to ponder: What exactly are we hoping to replicate? And what are we leaving behind?

Chomsky vs. Chomsky: First Encounter will be presented at the 2020 Sundance Film Festival – New Frontier from January 23 to February 2, 2020.

New Frontier at the Ray
1768 Park Avenue
Park City, UT
84060

Explore the potential and limitations of AI with a VR experience that brings you face to face with CHOMSKY_AI, an entity built from the digital traces of renowned thinker Noam Chomsky.

Artificial intelligence (AI) is everywhere—from the photo enhancer in your smartphone to Alexa, the virtual assistant in your kitchen. But what is AI? Aside from the constant promises and expectations of a better future, what does it mean to try to emulate human intelligence? The National Film Board of Canada, Schnellebuntebilder and EyeSteelFilm bring you Chomsky vs. Chomsky: First Encounter, a 10-minute prologue to a VR experience that lifts the curtain on machine intelligence and lets you peek inside.

The AI gold rush is steadily gaining ground in health care, business, education and almost every facet of our lives. But as big companies race to build machines that promise to “think like humans,” the technology itself remains opaque, hyped up and mythologized. We think we know AI, but it feels out of reach. This project seeks to raise AI literacy by informing us about its potential and limitations, all while reminding us of what’s special about human intelligence.

Chomsky vs. Chomsky aims to provide a playful and introspective interaction rather than an intellectual experience. Guided by no other than CHOMSKY_AI, our virtual host built from digital traces of Noam Chomsky, we’re invited to think about the quest for AI—what we’re told it can do and, perhaps above all, what we want it to be.

Noam Chomsky is a man who wears many hats: linguist, philosopher, cognitive scientist, social critic and political activist. Featured in thousands of videos, pictures and recordings on the Internet, he is one of the most digitized living intellectuals. But Chomsky has also devoted his entire life to understanding the human mind, making him a perfect guide for this journey.

Chomsky vs. Chomsky: First Encounter is directed by Sandra Rodriguez and produced by the NFB, Schnellebuntebilder and EyeSteelFilm. Michael Burk from Schnellebuntebilder and Kling Klang Klong share the visual and sound design credit. Cindy Bishop from M.I.T. Media Lab serves as technology lead and Guillaume Petitclair and Olivier Blais from MOOV.AI are the AI leads.

Chomsky vs. Chomsky: First Encounter is an immersive experience.

In designing it, the team used different AI systems, which were inspired by the tools gaining popularity in our everyday life. One such example is a simple chatbot, which allows you to “talk” to the AI system using speech-to-text technology. The reverse technology is used to allow CHOMSKY_AI to respond, combined with deepfake voice creation. There is also a tool for intent analysis, which allows the system to get an idea of what you’re asking about—whether it’s a question, a joke, or what subject it’s about. The team also designed a more complex conversational AI system, which learns and gets better as more information is fed to it. This is a system under development, and users help train it.

Using a VR headset, audiences will be free to ask questions, interact and speak with CHOMSKY_AI, following the common conventions of conversation. As in any real conversation, visitors may choose to move on to different subjects, while CHOMSKY_AI prefers to finish his thoughts on the current one. Although the conversation is free to go in any direction, it is still guided. Built from the arsenal of existing online material, the project demonstrates that no matter what we want to believe, all AI systems are still very much structured.

While Noam Chomsky was informed of the project, this experience was created strictly from material found online. In fact, one of the things that makes this so interesting is that Chomsky himself would likely disagree with the idea of being turned into an AI experience. Though this provides for much irony, it also ensures the project will stay true to its goal, as the material used critically assesses the way we, as a society, seem obsessed with using digital traces to emulate the human mind.

The user experience for this project is driven by conversation, so sound plays a crucial role. In speaking to CHOMSKY_AI, voice is integrated as an input into the system. Words used and their tonality affect the virtual world.

Chomsky vs. Chomsky: First Encounter is a prologue—a conversation starter. It was created using 4V (or vvvv), a platform that enables quick prototyping and developing. It’s a toolkit that enables general applications of real-time video synthesis and programming, allowing for the creation of an environment that can be used in VR. It facilitates the intersection of strong storytelling with interactive and real-time data-flow analysis, as well as the visual programming of audio and video.

AI is talked about everywhere. Hailed as an inevitable way of the future, the term encompasses a vast array of technology, some of it already part of our lives: from our relationships to Siri, to Alexa, to self-parking cars.

Behind the promises and expectations for AI, there is also much bluffing (and debate) about what it really is, and what it can or could do in the near future. On the one hand, this blurriness means opportunities: everything is open! On the other, we can already point fingers at questionable practices, like discriminatory surveillance and biased algorithms, and deepfake videos of politicians. How can we understand enough about AI to make sure we take part in steering its future?

Now is the time to have this conversation. After all, it’s the role of artists to question pragmatic and economic-driven futures. There is no better way to question a technology than by playing, disrupting and creating with it.

A general promise behind the latest developments in artificial intelligence is that it can emulate human behaviour. Many AI projects even go as far as to recreate famous people from the past—like an artificial Dali, or an AI-recreated Audrey Hepburn, or JFK and Nixon delivering deepfake speeches.

Chomsky vs. Chomsky: First Encounter lures visitors with a similar promise. What if we could replicate one of today’s most famous minds? What if you could meet and engage with CHOMSKY_AI?

Noam Chomsky, the famous linguist and scientist, has devoted his entire life to understanding the way our minds work. Known for his political statements, theories on intelligence, and hundreds of books and speeches, he has been interviewed on every subject over a period spanning almost 60 years, making him one of the most recorded and digitized living intellectuals. The pool of data we can use to create an AI system is thus extensive. What’s more, Chomsky is also known for his theory of natural language, an inspiration for “natural language processing” (an important backbone of AI). He becomes a perfect case-study and a perfect guide to help demystify AI.

But here’s the catch. Chomsky himself is skeptical of artificial intelligence. As he so often repeats, we know very little about how our minds work. So what exactly are we trying to replicate with AI?

Chomsky vs. Chomsky: First Encounter invites visitors to enter a virtual world, where we engage and interact with an AI system still under construction. CHOMSKY_AI is not a copy. It learns from a corpus of Noam Chomsky’s critical reflections. It builds and evolves from interactions, and constantly questions what it means for itself to exist. In an ironic twist, CHOMSKY_AI challenges us to consider: In this race to create machines that emulate who we are, what are the metaphors we’re willing to accept for ourselves? And what are we leaving behind?