Apple’s privacy pledge complicates its AI push


Cook’s steadfast aversion to the cloud presents a challenge as Apple tries to build up new features powered by machine learning and AI. To build and run machine learning services you need computing power and data, and the more you have of each the more powerful your software can be. The iPhone is beefy as mobile device goes, and it’s a good bet Apple will add dedicated hardware to support machine learning. But it’s tough for anything it puts in your hand to compete with a server—particularly one using Google’s custom machine learning chip.

I appreciate whatever complications Apple is going through in their stance on privacy. I much prefer their approach to these issues over Google’s and Facebook’s.

  • I also appreciate what Apple’s doing, but part of me wonders why they can’t use the cloud but discard your data afterwards. Like Duck Duck Go.

    • Kriztyan

      Because if it is written at any point, it can be captured, or reinstated. I am only guessing, but I think that it must be long those lines.

  • brisance

    Not this false narrative again.

    For machine learning, a large dataset is needed to TRAIN a model. In general, the larger the training dataset, the better. Once a model is TRAINED, you don’t need to use the training dataset anymore. Depending on how the model is set up, having more data may not necessarily benefit the model; it may even negatively affect the model’s accuracy. This is called over-fitting (high variance). See also curse of dimensionality.

  • It’s weird Wired would run this story consideirng just last August they ran a story explaining, from the horses’ mouths so to speak, how this is all going to be handled.

    Slow and steady might just win this race, and it definitely will help keep our privacy (or at least keep honest folks honest).

    • rick gregory

      Ha. Posted that, scrolled down, saw this… 🙂

      • I have the link in my Safari Read list so I can bring it up for threads. (Usually at other sites and on Facebook and such.)

  • rick gregory

    It would be nice if Wired’s writers read their own publication. This, last year, on differential privacy and Apple’s use of it: