Key Scientific Challenges Blog Series: Privacy and Security

Mar 15, 2010

On January 27 we announced the kick-off of our 2010 Key Scientific Challenges Program. Earlier this week we launched what we hope will be a thought-provoking series of guest blog posts here on Yodel Anecdotal that offer a quick overview of these scientific challenge areas. Check out our first post on green computing. Today’s entry is another incredibly important issue and difficult challenge posed by the Web – privacy and security. We’ve recruited Ashwin Machanavajjhala from Yahoo Labs to share his point of view on why privacy and security are Key Scientific Challenges. As a popular destination on the Web, millions of users interact with Yahoo every day. They search for information online, read the news, click on ads, upload, tag and share photos, and so on. Each of these billions of individual actions sheds some light on a particular user’s interests – what they like, what they want to know more about, buy or do in their spare time. They are all valuable cues that can be used to customize someone’s Web experience in a way that’s personally relevant to them. Ultimately, a personally relevant experience is at the root of what we’re trying to deliver with all of Yahoo’s products. The information that comes our way is a prime asset for making our users happy. But at the same time, with all that information comes a huge responsibility. We need to take extreme care to make sure people’s privacy is not breached. This responsibility to earn and keep our users’ trust is not just a matter of Yahoo policy (although that is critical too), it’s also a technical challenge that requires scientific innovation to continuously improve and maintain. When it comes to the trust of our users, we’ve learned lessons on how to approach both the policy and technical elements of privacy and security. On the policy side, we are extremely proud of our Data Anonymization Policy, which has received wide support and affirms our commitment to help protect our users’ privacy. Yahoo's policy both dramatically reduces the time we hold personal data and increases the scope of log data covered under the policy. Under the policy, we anonymize user log data, including deletion of total IP address, after 90 days with limited exceptions to fight fraud, secure systems and meet legal obligations. We’re also expanding our commitment to include data on page views, page clicks, ad views, and ad clicks as part of this policy. On the technical side, we’ve invested in giving our users the ability to understand and shape how we interpret what’s personally relevant to them by launching our Ad Interest Manager, a central place where Yahoo visitors can see a concise summary of their online activity and make easy, constructive choices about their exposure to interest-based advertising served by the Yahoo Ad Network. And our scientists at Yahoo Labs have also been active in academic privacy research, examining new mathematical definitions of user privacy and developing novel technologies for sharing and utilizing user information to improve Web experiences in a privacy-preserving manner. These experiments and research have been published widely, opening up new avenues of investigation on issues that are critical to Yahoo, like protecting search log data and dealing with so-called vanity queries, to new privacy frameworks that can be used in any field where making data available broadly is a key objective in overall innovation, as it is with the World Bank, Census Bureau and medical institutions publishing important public health studies, for example. These efforts are just the beginning, though, which is why privacy and security is a Key Scientific Challenge area, and we can’t wait to see what the word’s aspiring minds have to say on the topic. Ashwin Machanavajjhala Research Scientist Yahoo Labs

More information at:

  • Yodel Anecdotal