What does AI mean for consumer devices?

IFA 2017, Europe’s largest technology trade show, is soon to kick off in Berlin with an array of consumer electronics companies expected to show off their latest products and innovations. With artificial intelligence (AI) changing the way in which organizations innovate and communicate their processes, products and services, AI is expected to be a component in many of the announcements coming out of IFA.

We've asked Annette Zimmermann, research vice president at Gartner, to share her thoughts on what announcements in the area of AI-powered devices she expects in Berlin.

Q: As AI and virtual assistants gain popularity, what does AI mean for devices and smartphones in particular?

A: We are being inundated by the launch of new AI-powered devices or services each month, and it feels like AI has become the default marketing term for any technology product. We need to cut through the hype and take a closer look at products that are "AI-powered" to understand what machine learning methods are behind them. Many different techniques of machine learning exist, and we know that the type of data used for training machines has a direct impact on the quality of AI products.

Nevertheless, the era of intelligent devices is here. Currently, they are mainly in the form of virtual personal assistants (VPAs), but this will evolve quickly over the next two to three years. Personal coaching systems use machine-learning algorithms to enable a highly personalized user experience, for example. Ear-worn wearables such as Xperia™ Ear, Samsung Gear IconX and Apple AirPods can become a "daily assistant in your ear," but these little helpers are dependent on the smartphone when it comes to architecting such systems.

In the future, the question will not be whether AI can enhance our devices or smartphones, but how much information we should allow these systems to share in order to make our AI-powered smartphones deliver a rich and valuable customized experience. Do I allow my personal health app to share information with my fitness app so the information combined provides a much better overall picture of my health? Do I then allow this information to be shared and used by my coach to recommend the right fitness plan for me? Ultimately, it will be about connecting all the dots — the different data points such as emotional state, preferences, location data and behavior — and then training machines using this wealth of data to enhance the services that a device can deliver.

Q: On the topic of AI, what announcements are you expecting to hear at IFA?

A: I expect a fleet of product announcements, similar to previous years. Those related to AI will be centered on the integration of VPAs into new products and the enhancement of VPAs. Some connected-home products are expected to integrate Amazon Alexa, and Huawei is said to be working on an AI-enhanced service for its upcoming flagship product. Currently the market for VPA service on smartphones is dominated by Google Now/Assistant and Apple's Siri, while Samsung has taken a different approach with Bixby.

We will see more conversational AI systems from smaller vendors fiercely competing with larger vendors such as Microsoft and Apple. Needless to say, some large vendors have a clear advantage due to investments made in technology assets such as knowledge graphs. It is difficult for smaller vendors to match such investments, but we shouldn't underestimate them — many own large databases that are used to train machines.