Over the past month, we’ve seen a steady flow of articles discussing machine learning, online privacy and personalization. While each of these topics is interesting in and of itself, the intersection of the three is where the debate must take place. What data should machines be allowed to consume to provide consumers with a personalized experience—while not infringing on personal privacy? What role do humans play in determining how machines leverage data? The articles below present interesting perspectives that are part of the larger debate surrounding personalization and privacy.
“It’s Not Skynet Yet: In Machine Learning, There’s Still a Role for Humans,” GigaOM – While big data coupled with machine learning provides businesses with a powerful tool for delivering personalized online experiences, humans play an integral role in determining what data sources to feed machines when developing the algorithms that inform machine learning. At GigaOM Structure:Data earlier this month, Baynote’s Scott Brave pointed out that humans can “build intuitions and holistic pictures in their minds and see these connections that the machine might not even have the possibility of seeing because it doesn’t have the right data.” Sure, a machine may be able to process more data than an individual person, but we need to remember that the human role in defining how that data is processed is at the core of all actionable big data and machine learning.
“What You Can Learn about Online Shoppers by Watching Them” AllThingsD – When we talk about companies observing shopping behaviors, the conversation usually surrounds how retailers can use that information to deliver a personalized experience that does not infringe on a consumer’s privacy. But what if the data collected on a user can predict his or her intent to perform malicious acts on a website? Do the privacy implications change? This article by Liz Gannes explores how companies are now watching consumer shopping patterns to gain insights into who is most likely to commit fraud. For example, “People who shop between midnight and 1 a.m. are twice as likely to be fraudsters.” While this information is useful in identifying potential malicious acts, the vast majority of shoppers are law abiding citizens and that their personally identifiable information should be protected as such.
“Web Privacy Becomes a Business Imperative,” The New York Times – Since we’re on the subject of privacy, let’s look at an article by Somni Sengupta that discusses how companies are starting to take consumer privacy more seriously. While there has been little movement on privacy legislation, companies such as Microsoft have recently put greater emphasis on protecting consumer data. There are a couple potential reasons for this policy change. First, companies that act independently may be trying to stave of Federal legislation that can be unpredictable and prescriptive. Alternatively, these companies may be reacting to consumer sentiment around the desire for more transparent data collection and use policies. Whatever the cause of the increased attention to consumer privacy, I believe this conversation must continue to take place both within companies and among consumers.
“Rajuten: The Biggest eCommerce Site you haven’t Heard Of” CNN Money – This article on Japan’s eCommerce leader, Rakuten, introduces a principle I am particularly fond of: “Omotenashi,” or the concept of providing high quality and personal service. “In Japan, if you go to a local coffee shop, or a grocer or any local business, you will find a very high level of personal service,” says the company’s CEO, Hiroshi Mikitani. “The internet should be that way as well.” At Baynote, we full heartedly agree with Mikitani, but understand that personalization must occur against the privacy backdrop discussed above. The real trick is providing a highly personalized online experience while respecting consumer privacy by only using consumer information to benefit the end user and not third parties.