Skip to Content

Project Oxygen’s New Wind

Look out, PCs. MIT’s ubiquitous computing effort is taking technology out of the box.
December 20, 2001

Michael Dertouzos, late director of MIT’s famed Laboratory for Computer Science, had a vision: pervasive, human-centered computing. “We find ourselves in the junction of two interrelated challenges: Going after the best, most exciting forefront technology; and ensuring that it truly serves human needs,” he wrote in the lab’s mission statement. Last year, the LCS-together with MIT’s Artificial Intelligence Laboratory-launched an ambitious effort to build that vision, called Project Oxygen.

An umbrella for more than 30 faculty members, Oxygen supports research aimed at replacing the PC with ubiquitous-often invisible-computing machines. Projects run a gamut from video recognition to nomadic networking to chip design. The Defense Advanced Research Projects Agency contributed $20 million; a consortium of private companies-including the Acer Group, Philips, Delta Electronics, Hewlett-Packard, NTT and Nokia-ponied up thirty.

In August, just as Project Oxygen began to bear fruit, Dertouzos died after a long illness. Carrying on his vision-and advancing their own-are his colleagues and intellectual heirs, among them LCS associate director Anant Agarwal, MIT Artificial Intelligence Laboratory director Rodney Brooks and LCS director Victor Zue. Last month, Brooks and Zue (Agarwal was unavailable) sat down to share that vision with technologyreview.com writer Eric Brown.

TR: How has the death of Michael Dertouzos changed the project’s vision?

ZUE: If you ask us what Oxygen is all about, we’re all going to be like blind men describing an elephant. We all have different ways of looking at it. Michael’s bumper sticker was “Doing more by doing less.” Anant [Agarwal] is saying that in ten years computing and communications are going to be so abundant that they will be free, so what are we going to do to make use of it? Rodney would say that computers ought to enter the human world rather than the other way around. And I would say we should pay more attention to semantics and intent rather than syntax and form.

TR: Can you give an example of what you mean by “semantics and intent”?

ZUE: There are many examples. In speech-based interfaces, it is important that we understand what people say rather than simply transcribe. Without understanding the meaning of their queries, we won’t be able to satisfy their needs. In terms of mobile networks, we need to move from specifying the precise IP address of devices to something more functional and intentional like “the nearest uncongested color printer.”

TR: Project Oxygen conjures a vision of computing as pervasive as the air we breathe. A neat idea-but what problem does it solve?

ZUE: Oxygen tries to ensure that the technology and artifacts are human-centered by attending to their needs and wants in a way convenient to them rather than to the computers. We want to radically change the way humans deal with their information-related activities. Pervasiveness is but half the ingredient of this complex solution.

TR: Elsewhere, you’ve described two ways to interact with Project Oxygen: a handheld device you call H21, and an environmental system you call E21, or the “intelligent room” which consists of speech, video and motion detectors. How far along are these ideas?

BROOKS: We have about a half dozen offices with E21s.

ZUE: There are about ten H21 prototypes, each with slightly different capabilities.

TR: A gaggle of handheld makers are pursuing projects similar to the H21. Why compete with them?

BROOKS: It is a delicate balance. We’d be foolish not to rely on what’s happening in the commercial world, but we’re trying to do things a little beyond what they’re doing, so we can’t wait for them. For instance, the camera that we have on the H21 adapts to the brightness level so we get orders of magnitude more brightness variation. This is important. You’re holding a handheld unit and it’s looking up at you. It’s seeing a dark face with bright lights behind it, so you need an adaptive camera.

TR: Besides speakers and display screens, how will Oxygen devices “talk back” to the user?

ZUE: There are many different things in that area. It could be a shape or a fabric. [We’re] working on new rendering and display technologies for very large, wall-sized displays so that we’ll be able to construct a truly three-dimensional dynamic display. Anant Agarwal’s work on the RAW microprocessor will ultimately produce computational fabrics that can serve as display elements.

TR: Where has the research made the most progress?

ZUE: We’ve made surprisingly fast progress in networking. For example, Migrate is an architecture for vertical host mobility. The neat thing is that you can change network protocols from home to car to office, all without involving a third party. It uses dynamic updates to the DNS to track host location. Existing connections are retained using connection migration that enables connections to negotiate a change in endpoint IP addresses.

TR: One of the most interesting aspects of Oxygen is also one that users might object to the most: continual surveillance by microphones and video cameras. Are the advantages of these technologies enough to overcome privacy concerns?

BROOKS: Would you like to have a microphone on in your bedroom that listens to everything you say 24 hours a day? Well you have one. When you hang up a modern telephone it doesn’t physically disconnect the microphone. So you have to be a little sophisticated about how you’re going to use the thing. We have to be careful that we provide those sorts of protections. There has to be full disclosure. You can’t have a person being videotaped without them knowing it.

ZUE: Privacy is one of four interlocking issues that we must address in a pervasive, human-centered world. The first is nomadicity: people and devices are going to move around a lot, so we must provide location-aware support. But location tracking is not good. There are a whole bunch of scalability and privacy issues; people don’t want to be tracked. The second issue is that devices must be able to maintain anonymity; otherwise we will all have to carry assorted gizmos with us, making us look like road warriors. The third is personalization; we must be able to transform and customize anonymous devices to suit our needs so information can follow us around. Last but not least is the issue of security and privacy: we must make sure that personalization does not result in the invasion of privacy.

TR: Can you give an example of how Oxygen protects your privacy?

ZUE: I can pick up an anonymous H21 in someone else’s office, and it will recognize who I am and fetch information that I need. When I say, “call home,” it will turn itself into a cell phone and dial the right number. But when I’m done, it will promptly forget everything about me and return to its anonymous state.

TR: One of Oxygen’s goals is eliminate the drudgery of human/computer interaction, by, for example, monitoring users to guess what they want to do next. But isn’t it possible that overriding the computer’s bad guesses could take more work than simply issuing the command yourself?

BROOKS: How much control you give the machine and how much you interact with it is a serious issue. It may be painful. Michael Dertouzos loved customizing things. Others don’t.

TR: What surprises have you encountered testing the Oxygen prototypes?

ZUE: We’re always finding surprises.

BROOKS: One surprise was that nobody wanted to wear headsets. So we realized we had to do something about microphone arrays. In the Intelligent Room, we’re combining arrays with personal tracking technology using video. We’re even looking at incorporating lip motion recognition.

TR: When an intelligent room gets crowded, how does the computer know who to pay attention to?

ZUE: We are trying to combine speech and vision in ways that they can complement each other. In a very noisy environment you invariably begin to pay attention to people in terms of their facial expressions. Lip reading can improve speech recognition performance. We might also be able to steer the microphone array toward the person whose mouth is moving.

TR: The personal computer has come to be equated with the concept of work. But if Oxygen kills the PC, what is the new work metaphor? Does Oxygen change the nature of work?

BROOKS: It’s already changed over the course of the last few years. Now we’re moving into the domain where there are many more computers than there were five years ago. When people in this building started developing time-sharing systems [in the 1960s], one of the surprises that came out of that was e-mail. That wasn’t a stated goal. The hackers who were building the systems wanted to leave messages for each other because they were working around the clock. I expect as we’re working on these things we’re going to see new ways of working that we haven’t thought of yet.

Keep Reading

Most Popular

Large language models can do jaw-dropping things. But nobody knows exactly why.

And that's a problem. Figuring it out is one of the biggest scientific puzzles of our time and a crucial step towards controlling more powerful future models.

OpenAI teases an amazing new generative video model called Sora

The firm is sharing Sora with a small group of safety testers but the rest of us will have to wait to learn more.

Google’s Gemini is now in everything. Here’s how you can try it out.

Gmail, Docs, and more will now come with Gemini baked in. But Europeans will have to wait before they can download the app.

This baby with a head camera helped teach an AI how kids learn language

A neural network trained on the experiences of a single young child managed to learn one of the core components of language: how to match words to the objects they represent.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at customer-service@technologyreview.com with a list of newsletters you’d like to receive.