Click here to watch the TEDTalk that inspired this post.
My first reaction to Jennifer Golbeck's TED Talk on the "Curly Fry Conundrum" was "Curly Fries! You've got to be kidding me!" Computers capable of establishing a link between high intelligence and liking a Facebook page on Curly Fries sounded like a classic case of over-fitting the data, of finding correlations between two sets of data that weren't connected at all. We've all heard of bizarre correlations like the Super Bowl Indicator, which claims to predict the future performance of the stock market based on which team wins the Super Bowl.
But the more you think about it, the more it seems likely that tomorrow's super-powerful computers will be capable of using algorithms to pick the proverbial needle out of the digital haystack. In short, computers will know us better than we know ourselves, thanks to the increasingly voluminous trail of data that we leave behind us every day. They will know our political preferences and our unexpressed longings. These future computer algorithms will be more powerful than the Target algorithms referenced by Golbeck - computers that were capable of predicting that a teenage girl was pregnant based on running the data on a vast number of purchasing decisions and online behaviors.
Which raises a number of interesting philosophical questions. To what extent, for example, is our digital "self" different from our real-world, physical "self"? And, if our digital self is the same as our physical self, will computers one day be capable of carrying on relationships with humans just by processing our personality as a string of 1's and 0's? Will computers be capable of an emotional response if they recognize when we are going through a difficult period in our lives? What if our trail of data, for example, suggests that we are experiencing a bout of depression after the end of a relationship?
As Golbeck suggests, there are a lot of people out there who are increasingly concerned about who has our data and what they are doing with it. Those are probably the same people who use avatars on social media instead of their real images, or the people who try to make their lives online as anonymous as possible. But, most likely, most of us simply find it too tiring, too complex, to pay much attention to all the privacy settings out there. How many of us, for example, actually change the password settings when we are supposed to? We assume, naively, that there must be some kind of law out there that keeps corporations from going too far with all that data they are collecting on us.
That being said, it would be great to see some kind of innovation that solves the Curly Fry Conundrum. One idea that strikes me as incredibly innovative is thinking of our personal data as a form of "human capital" that we all possess. And that human capital should be earning a return, just like other forms of capital. In short, if our data is that valuable to marketers and other companies online, they should be willing to pay us for that data. The World Economic Forum in Davos has even discussed plans to create a separate asset class around data, in which people invest in people's data, the same way they might invest in securitized assets like mortgages.
Which sounds a bit crazy until you start thinking about the ways that we already invest in human capital in wonderful creative ways without thinking of it in those terms. It's baseball season right now, for example. How many people are involved in fantasy sports leagues where they buy and trade athletes based on data and performance? We've somehow abstracted these professional athletes into 1's and 0's, no longer viewing them as real-world personalities, only as data and statistics that change on a daily, weekly and monthly basis. That might just be a foreshadowing of our personal data future.
We want to know what you think. Join the discussion by posting a comment below or tweeting #TEDWeekends. Interested in blogging for a future edition of TED Weekends? Email us at tedweekends@huffingtonpost.com.
My first reaction to Jennifer Golbeck's TED Talk on the "Curly Fry Conundrum" was "Curly Fries! You've got to be kidding me!" Computers capable of establishing a link between high intelligence and liking a Facebook page on Curly Fries sounded like a classic case of over-fitting the data, of finding correlations between two sets of data that weren't connected at all. We've all heard of bizarre correlations like the Super Bowl Indicator, which claims to predict the future performance of the stock market based on which team wins the Super Bowl.
But the more you think about it, the more it seems likely that tomorrow's super-powerful computers will be capable of using algorithms to pick the proverbial needle out of the digital haystack. In short, computers will know us better than we know ourselves, thanks to the increasingly voluminous trail of data that we leave behind us every day. They will know our political preferences and our unexpressed longings. These future computer algorithms will be more powerful than the Target algorithms referenced by Golbeck - computers that were capable of predicting that a teenage girl was pregnant based on running the data on a vast number of purchasing decisions and online behaviors.
Which raises a number of interesting philosophical questions. To what extent, for example, is our digital "self" different from our real-world, physical "self"? And, if our digital self is the same as our physical self, will computers one day be capable of carrying on relationships with humans just by processing our personality as a string of 1's and 0's? Will computers be capable of an emotional response if they recognize when we are going through a difficult period in our lives? What if our trail of data, for example, suggests that we are experiencing a bout of depression after the end of a relationship?
As Golbeck suggests, there are a lot of people out there who are increasingly concerned about who has our data and what they are doing with it. Those are probably the same people who use avatars on social media instead of their real images, or the people who try to make their lives online as anonymous as possible. But, most likely, most of us simply find it too tiring, too complex, to pay much attention to all the privacy settings out there. How many of us, for example, actually change the password settings when we are supposed to? We assume, naively, that there must be some kind of law out there that keeps corporations from going too far with all that data they are collecting on us.
That being said, it would be great to see some kind of innovation that solves the Curly Fry Conundrum. One idea that strikes me as incredibly innovative is thinking of our personal data as a form of "human capital" that we all possess. And that human capital should be earning a return, just like other forms of capital. In short, if our data is that valuable to marketers and other companies online, they should be willing to pay us for that data. The World Economic Forum in Davos has even discussed plans to create a separate asset class around data, in which people invest in people's data, the same way they might invest in securitized assets like mortgages.
Which sounds a bit crazy until you start thinking about the ways that we already invest in human capital in wonderful creative ways without thinking of it in those terms. It's baseball season right now, for example. How many people are involved in fantasy sports leagues where they buy and trade athletes based on data and performance? We've somehow abstracted these professional athletes into 1's and 0's, no longer viewing them as real-world personalities, only as data and statistics that change on a daily, weekly and monthly basis. That might just be a foreshadowing of our personal data future.
We want to know what you think. Join the discussion by posting a comment below or tweeting #TEDWeekends. Interested in blogging for a future edition of TED Weekends? Email us at tedweekends@huffingtonpost.com.