5 Things Apple Could Do With Machine Learning Startup Laserlike
News broke yesterday that Apple has added another machine learning startup to its portfolio of smaller acquisitions, with this one focused on artificial intelligence that seems specifically designed for providing user-focused recommendations for everything from news, sports, and music to users' web browsing habits.
While it's fairly obvious what more privacy-invasive companies like Google or Facebook could do with Laserlike's technology, Apple's acquisition of Laserlike — and the company's overall efforts to look at the long game for Siri and other AI features — has gotten us thinking about areas in which the machine learning startup could be folded into Apple's new "Machine Learning and AI Strategy" division. Continue reading to learn 5 Things Apple Could Do With Machine Learning Startup Laserlike.
Apple News Recommendations
This one is of course the most obvious fit for Laserlike. The company previously had published an award-winning smartphone app that was specifically designed to let users search out and follow news, music, and sports topics. Laserlike described its app as using AI to provide "an amazing personalized lens of the world’s firehose of information."
With Apple on the cusp of launching a premium news and magazine subscription service, the ability to use Laserlike's technology for an AI-based content search and recommendations engine seems like natural synergy. Although Apple has traditionally focused on human-curated news — and will probably continue to do so in terms of what's generally recommended and fronted on its News service — more advanced AI would be invaluable in providing a much more personalized experience for Apple News, which presently relies on a fairly simplistic scoring engine.
Apple Music Playlists
Apple Music would be another clear place where Laserlike's AI could enhance Apple's recommendations engine, and like Apple News could enhance and build upon Apple's current human curation strategy, rather than supplanting it entirely.
To be fair, Apple already does a pretty good job of surfacing music content in its various "For You" recommendations, creating mixes of songs based on users' favourites, what their friends are listening to, and even recommendations for new music and relaxation. Still, while Apple's biggest advantage in this area is its use of human curation, rather than algorithms, the actual recommendations engine seems somewhat basic in simply factoring in a users likes and dislikes. Machine learning has the potential to greatly expand this, while Apple can still do it in a way that respects a user's privacy.
Siri Personalization
There's no doubt that Siri needs help, so it seems likely that at least some of Laserlike's AI skills will be used to improve the voice assistant. While there are a lot of areas for improvement, it seems that what the Laserlike team could bring to the table would be improved personalization.
Using Laserlike's technology, Siri could learn more about users' habits and routines, allowing it to respond in more personalized and relevant ways. While Apple has already been tackling this with its "Proactive Assistant" that was introduced in iOS 9, there's no doubt in our minds that the assistant could do to get a bit more intelligent when it comes to analyzing behavioural patterns, and particularly doing it more reliably.
Mail and Message Filtering
Despite repeated predictions of the demise of email, there's no shortage of third-party apps that attempt to do intelligent things to help users manage their inboxes. However, one thing that almost all of these apps have in common is the need to rely on a back-end service to handle the analysis, with all of the privacy implications that come with it.
For many users, email, text messaging, and photos make up the most private data on their iPhone, and even Facebook is starting to get this. Apple's focus on putting its machine-learning directly into its own silicon — something it's already famously done with its Photos analysis — is a huge selling point for those concerned about privacy.
Although some will argue that Apple's Mail and Messages apps should remain fairly simple, Apple should at least provide some intelligent on-device filtering for things like notifications. Apple's VIP List and Thread Notifications are both useful features, but they're beginning to seem like an anachronism in a day when many other apps can much more intelligently surface those messages you need to see, whether it's a new message coming in, or an old one that you've forgotten to reply to. There's no reason why Apple shouldn't be able to do both of these things, and do them directly on an iPhone, iPad, or Mac, providing AI that's independent of any particular email provider.
Health and Fitness Analysis
Another area where Apple could benefit from better machine learning analysis is in both health and fitness — two areas where the company has put in some rudimentary intelligence, but hasn't really leveraged true AI technology quite yet. For example, an Apple Watch will provide automatic workout detection, but this uses simple motion detection rather than factoring in any kind of other information such as user habits and schedules.
More importantly, however, Apple's HealthKit framework is essentially just a repository for data at this point — it records a lot of information, but Apple itself isn't really doing anything with this information other than storing it. Advanced machine learning could allow Apple to provide recommendations based on this health data that could turn a user's iPhone and Apple Watch into a virtual healthcare assistant, factoring in all of a user's health points to suggest nutritional tips, exercise regimens, water intake, and other positive lifestyle initiatives.
With Apple's Health Records feature, machine learning could also be used to analyze the wealth of health data recorded by Apple Watch users to provide primary care providers like doctors with intelligent snapshots of a patients relevant health data — something that at least one analyst suggests doctors might be willing to pay for.