Following Apple’s keynote at the Worldwide Developers’ Conference, WWDC16, last month, it seems Apple reinforced its position as a key defender of privacy. Making it more difficult for hackers, the US government and Apple itself to access user data and putting themselves up against data hungry companies. However, Apple still do need some access to user data to be able to offer the best services and that’s where the concept of differential privacy comes into play.
“Differential privacy is a research topic in the area of statistics and data analytics that uses hashing, sub-sampling and noise injection to enable this kind of crowdsourced learning while keeping the information of each individual user completely private,” Craig Federighi, Apple’s SVP of Software Engineering, explained.
Still not sure what differential privacy means? Here’s a clearer explanation:
Differential privacy works by algorithmically scrambling individual user data so that it cannot be traced back to the individual and then analyzes the data in bulk for large-scale trend patterns. The goal is to protect the user’s identity and the specifics of their data while still extracting some general information to propel machine learning.
Differential privacy aims to provide the means of maximizing accuracy of queries from statistical databases while minimizing the chances of identifying the individuals from which the records were built.
In order for this concept to work there must be some trade-off between the accuracy of the statistics estimated in a privacy-preserving manner, and the privacy of the individuals.
In the case of Apple, their trade-off consists in mixing on-device data with noise in order to obscure personal information, guaranteeing protection of individual identities while also enabling actionable insights into what they do.
Apple says that “Differential Privacy adds mathematical noise to a small sample of the individual’s usage pattern. As more people share the same pattern, general patterns begin to emerge, which can inform and enhance the user experience.”
Apple is using differential privacy to enable insights in four specific ways:
- New words added to local dictionaries;
- Emoji words (e.g. pizza) typed by the user where iOS can suggest emoji replacements;
- Spotlight deep links used inside apps;
- Lookup hints within Notes.
For these different sections, Apple gives its users the choice to opt out. If instead they would like to opt in, Apple guarantees that their privacy will not be compromised.
Your users will probably feel more comfortable sharing their information within this privacy-friendly environment since the commitment exists not to misuse their personal information as it will be non-identifiable. So what if differential privacy could be used more extensively? How could you implement differential privacy in your business?
- Data-driven product development (which features do users use, have problems with, etc.) by gaining further insights into user behaviour
- Recommended (related) products/services based on what other users are looking at (but maybe not even purchasing)
- Suggested search results or answers to FAQs
- Lower chances of being hacked and the consequences of a potential data breach by anonymizing part of your dataset
- And much more.
As the giants are leaning towards more privacy-friendly services, will you follow their lead? What are you proactively doing to be innovative in the privacy field? Are you integrating privacy by design into your big data strategy?
To learn more about the latest trends around privacy in mobile, download our white paper on the Top 5 Privacy Trends in Mobile 2016 or contact us.
Senior Counsel, International Operations and Privacy Specialist