MENLO PARK, Calif.--The first real question about security and privacy emerged more than an hour into an expert panel on the future of human-computing interfaces (HCI). That could be a function of the audience makeup (technologists) or the fact that as we journey deeper into the land of ubiquitous computing, Internet of Things or whatever you want to call it, we, as a society, are starting to come to grips with privacy and security's role in the human-computing interface.
Bill Mark from SRI, Walt Johnson from PARC and Andy Wilson from Microsoft spent Tuesday evening (Aug. 27) fielding questions from newspaper reporter James Temple and the audience at a Churchill Club event here at SRI's headquarters.
The panelists roamed widely over the nature of anticipation, voice control, multi-user contention and mediation, ubiquitous computing and the Internet of Things, Google Now and other topics. All of these technologies require varying amounts of user data, uploaded to the cloud, to function properly, but it wasn't until the program was nearly over when audience questions turned to the nature of privacy. Five years ago, that question would have come up a lot sooner.
If society comes to grips with security and privacy issues, then the next several years are going to be fascinating. Take, for instance, the notion of anticipation: computing devices so connected and sensor-smart that they anticipate your needs.
"Imagine a system knows not only whether you opened an email but how long you looked at it and which lines you read," said SRI's Mark, the company's vice president of Information and Computing Sciences. "So if one of your teammates is assuming or not assuming you read that email, it can act differently for you."
You should be able to sit in a meeting with a device with anticipatory functionality and be served a meeting summary or relevant documents to use during the meeting, he added.
"That's part of the future of human-computer interaction," Mark said.
Microsoft's Wilson said it's curious that as a culture we've invested lots of time, money and effort into researching these technologies and developing prototypes and yet we still have whiteboards in our offices.
It's hard to develop the right replacement, he said. "It's tempting to add this, that and the other feature and pretty soon you have drop down menus," said Wilson, who is a principal researcher with Microsoft. "If you can't be engaged within seconds you're not going to use it."
You lookin' at me??
One non-trivial challenge is precision in many of these conceived systems. Take lights for instance. Wilson said it's relatively straight forward to turn on your house lights, even after you stumble home after a few cocktails.
But in an IoT world, electronics design is much more challenging. For instance, people in test groups favor using their voice to control lights in the future. But there isn't enough technical precision for a light across the room to understand--through tracking your eye movements for example--that you're talking to it.
"We don't have a model that allows us to figure out what needs to be selected and (then) extract only the information at the level of precision that's appropriate to complete that task. You still have to get the mouse in the rectangle of the button no matter if there are only two buttons that are on the screen. That seems fundamentally wrong."
What's going to take off and what's not in the near term?
Voice recognition is hot, according to Mark, who noted that venture capitalists still knock on SRI's door to get the SRI-invented Siri voice-recognition technology for use in various ways.
Johnson, vice president of the intelligent systems lab at PARC, said that PARC is doing work in smart manufacturing where a user could take any 3D CAD design, drop it on a box and discover instantly whether it's manufacturable. If it is, the system lists the supply-chain options, parts requirements and their costs, and who could do the design.
"It starts to put information at the fingertips of the designer. Is it a good design? Is it sustainable. You're also democratizing who could do the design. There are 300,000 small manufacturers who don't use technology out there."
Old v. New
How will this research and development affect tried-and-true interface modalities, such as the keyboard?
Mark said the keyboard may fade but it likely will never disappear:
"There are some situations where the keyboard is just the right thing. If you think about a pen or something like a pen...it's been around for a really long time. That's because it works really well in certain situations."
If old technologies linger, so too do old mindsets. Microsoft's Wilson said right now "we're stuck in an event-driven model." One of the beauties of the computer mouse design is that it does nothing when you move your hand away. He added:
"It's either on or off. In the future, programming models are going to more like a probabilistic network. So the challenge is how do you develop using those tools?"
Society has largely adapted with the times, the panelists agreed. Johnson noted that Xerox in the 1980s had voice-controlled copiers that never caught fire in the market because users were uncomfortable talking to a machine. Today, few people think twice about barking orders into their phone or their car's dashboard.
What will drive adoption of these technologies, however, will be less the technology (we can do a lot of this already) and more the business models.
A lot of today's devices don't play nicely together technologically because they are walled gardens competing for your time, attention and dollars.
"The main reason information about you is Balkanized is because there are different business interests vying for it," said Mark. "Google wants to own this piece of (your) world; the phone provider wants that part of the world. Getting a business model where those things are share is difficult."
This evolution of society with its technology may be one reason the privacy question didn't emerge until late in the discussion.
Mark offered a story to illustrate that social evolution. His brother spent time working in a Cambodian village. After a day or two, he noticed a couple of guys on scooters waiting for him outside his workplace to offer him a ride. At first, he was queasy that suddenly the village knew about his habits and whereabouts, but he quickly grew used to having a waiting mode of transportation.
That's similar to what's happening now in digital networks.
"You want to share info with friends and family so they can anticipate your needs and do things for you. Unfortunately, we're now in an environment where you're not sure who all is seeing" your information.
Such were the visions of human-computing interface in 2020 and beyond. One complete unknown, curiously, was what the TV of the future looks like.
"I really don't know. Is it remote controlled? Speech driven?" said Wilson.
Mark pointed out that at the end of the day, the objective, for him, is not about improving human-computing interfaces.
"The goal is to enhance human life," he said. "I don't care how we interact with computers."
--Scaling the Semiconductor Memory Wall
--Circuit Design Trends: 3D-ICs Extend Scaling