Have you ever found yourself using your computer and thinking “No! That’s not what I meant. Isn’t it obvious what I’m trying to do?”
Today you can use your computer for an ever-increasing number of activities – planning a holiday, reading the news, creating music, chatting, shopping, budgeting or just satisfying an idle curiosity.
But there is a problem; your computer is fundamentally stupid. You have to tell it exactly what you want. Often you have to enter information many times in different ways. The computer has no understanding of you as an individual, so it must ask for your address and billing information for every online purchase. It has no understanding of the context of your request, so it can’t know when you type “Java” into a search engine whether you are at that moment interested in the programming language, the island, or the coffee.
The only way for computers to get smarter at this is for them to learn more about you. Fortunately, a number of companies are now building software that shows that if you allow your computer to watch and learn from you, it can become far more helpful.
Social agents, or Virtual Personal Assistants (VPAs) are a new type of software that observes what you do, and uses that information to anticipate our needs and act independently on our behalf. Software agents have been around for a while, but the concept is only now being applied to the social web. Here are some examples of social agent software:
- RescueTime monitors which programs you use and which websites you visit, allowing you to look at how you spend your time and improve your productivity.
- Twitchboard monitors your Twitter feed and automatically publishes your posted links to delicious, improving the quality of delicious’s repository for others, and providing you easier access back to your links.
- Google Latitude automatically publishes your location from your mobile phone, making it easier for friends to find each other. (Foursquare does this too, but you have to share your location manually).
- Infovark runs on your PC, monitoring your email, and file accesses, making it easy for you to access related documents, emails and experts in context, while you work.
- Dropbox, which I covered in a previous post, silently backs up your files while you work, and lets your colleagues know what you are working on.
- SmallBlue monitors your e-mails and chats, and builds a social network graph of an organization, which you can use to find out a “route of introduction” to an expert in a given topic.
- Aardvark is a social search engine. By sharing your areas of expertise it can help you find answers to questions that are much better answered by people than computers (such as recommendations or advice). It automates the process of asking around until you get an answer, and is able to ask more people than you could encounter alone.
- WebMynd monitors your web browsing, allowing you to easily rediscover sites you have previously visited and help others to find them.
- Siri is a virtual assistant for iPhone, which evolved out of DARPA’s CALO research project. Using a knowledge of your location, preferences and your past behaviour, it can automate tasks like booking a restaurant, checking opening hours or finding out what’s on near you.
Privacy and Trust
Most of us don’t trust computers, because we can’t see what they are doing. To make matters worse, many software companies capture and harness data about us for their own benefit – marketing, profiling or upselling – and they often do this secretly. If social agents are to succeed, we’re going to need a huge mindset-shift.
For this to happen, we will have to become skilled at knowing which software is recording what information about us, where it is storing it and what exactly it will do with that data. Right now, there is no legal requirement upon companies to make this easier for us. There are no technical tools to help us make these kind of judgements. And it is rare to be able to maintain legal ownership of your data once it lives in the cloud.
But these changes are vital if the social agent revolution to succeed. Right now the only people embracing these technologies are the most technologically-minded, because they are best equipped to make these judgements. But in future new laws and tools will enable all computer users to confidently harness software agents.
As we go forward, we shouldn’t be scared away from the idea of trusting our computers more and sharing more information. This is the only way that we can get to a future where computers can understand us without us having to change our natural behaviour or think like a computer. If we continue to distrust computers we will always have to work hard to convey what we want.
That’s not to say that we should embrace this change blindly; every new technology brings both benefits and risks. We will need to be careful about what we share and with whom, and continue to fight for data ownership, net neutrality and freedom of information, and against dubious laws such as the US Patriot Act and technologies like CView and Phorm.
So, next time your computer acts stupid, remember, you can help it get smarter. You just have to start sharing.
I recently dug into some other aspects of this topic and some of the challenges ahead in my 10 minute presentation at Bitnorth entitled “The Importance of Persona: Why your computer needs to know which hat you’re wearing” (slides & MP3 available via the link).
“The world’s biggest webcam?” by yakobusan on Flickr (CC)