2001: A Space Odyssey introduced us to artificial intelligence gone wrong. Ever since then, self-aware technology has been part of mass culture, from science fiction to horror to even an episode of the X-files. It's become cliché. But, however much we try and shake it away as the 21st century boogeyman, each of us wonders deep down inside, could that really happen?
In part two of this three-part series, I am going to be looking at is this notion that future interfaces will be more intelligent.
Artificial intelligence isn't something we should be totally scared about. It is already part of our daily lives. Amazon guesses what products you'd like to buy and those are the first things you see when you visit the site. The primary way Netflix displays ratings is not by what everyone else things, but what they think you'd think based on what everyone else thinks.
Real artificial intelligence is based on watching user behavior over time, watching community behavior over time, and guessing the preferred result to an action. It's not that different than the way any living creature learns over time, only that we have the option to change that action into a non-preferred result. So until Google starts sending me bad results on purpose, because they are holding a grudge that I'm still using Hotmail, I don't think we have anything to worry about
Science Fact
What we will be seeing is more intelligent recommendations based on two factors. Number one, semantic web will work to catalog the vast amounts of information out there in a more usable way. Number two, machines will come to recognize the subtleties of natural language. The intelligent interface will make it easy for me to just straight ask my computer, "Hey, what e-mail did I just look at in the last week about Aunt Rita's boat trip?" and get back some usable results that match a query I didn't even have to think about phrasing more search friendly.
Science Fiction
What we won't be seeing at least outside of the movies is HAL 9000. For the everyday user, there is no real benefit to creating a search engine that factors its own attitude into the way it delivers search results. Instead, that personality and that attitude will come from the users that create that content, rate that content, and share that content. And if you aren't already convinced that we won't be seeing a character-based interface agent in the future, just think about how much we all hate the Paperclip.
Where It's Poking Through
We know that Freebase and Twine are semantic applications, but what else is out there. The problem is that most good semantic web and artificial intelligence goes on behind the scenes. The idea is that intelligent agents are working behind the scenes, connecting and analyzing data, and the human comes in for the final recommendation. Nevertheless, I tried to pick some interesting examples.
1. Since Aurora was released by Adaptive Path, you better believe it featured semantic based tools. This first demo video features instances of that "behind-the-scenes" intelligent agent:
2. Microsoft adds Natural Language search into Vista. Before they turned this feature on "pictures taken last week" would yield zero results and now it actually yields, you guessed it, pictures you took last week:
3. Freebase Parallax is a user interface that details a method of finding information on the web beyond Google and Wikipedia. It allows you to aggregate and compare facts, and map them out on a visual timeline:
4. Nova Spivack writes about the future of the desktop. He claims that the smart desktop will help users become more productive, by adjusting interests, measuring current location, and keeping up with that changing social graph of ours.
"When reading something while you are taking a trip to Milan it may organize itself to be more contextually relevant to that time, place and context. When you later return home to San Francisco it will automatically adapt and shift to your home context. When you do a lot of searches about a certain product it will realize your context and intent has to do with that product and will adapt to help you with that activity for a while, until your behavior changes."
5. Then again, who cares about having the machine be intelligent for you, when you can just hook your brain up as an input. BrainLoop enables users to manipulate objects on the screen with purely by imagining movement, without single touch or click:
Where Social Will Fit In
So what does it mean for social features in the context of the intelligent interface? The biggest key is that it isn't just our data. A big component of human learning is not just by watching what people do, it is watching what people do with each other. The intelligent interface will understand who are friends are, what our relationships are centered around, and how our social graph changes over time.
When I access my wine information, the interface will understand that the most common friends that I discuss wine might not be the people in my local social circle, but the people I follow on Twitter. My options might change to allow me to more readily access wine note sharing via my blog, instead of something more localized, like scheduling an event in my area.
Tomorrow: Future Interfaces Will Be Everywhere
Will my t-shirt read me the weather? Will advertisements actually know my name on the subway? Read tomorrow to find out...
Marta Strickland
Strategists & Planners
Comments (1)
Again, love this series, it really helps the reader to look to the future, the possibilities and realities in a realistic fashion. Nice to break away from the 'preaching' or 'teaching' aspect of blogging.
Posted on September 10, 2008 06:43