Digital Assistants: The Quest for Artificial Intelligence

This month Apple signaled its commitment to advance the development of Artificial Intelligence (AI) in its products with the acquisition of tech startup Turi, a machine learning platform for developers. The move follows a growing trend by major tech companies such as IBM and Google to create and implement more advanced technology in the growing Chatbot industry.

For those new to the term, a ‘bot’ is simply software that’s been designed to automate tasks. Chatbots (from Chatter + Robot) take this to next level by simulating human conversation, making it seem as though you are talking with a human as you use the bot. While programs that recognize and respond to language have been around for decades, they have become a standard offering in consumer devices in more recent years. These programs use natural language processing, the ability to understand and use human conversational phrases, to interact with humans on simple tasks. Nearly all users of smartphones or messaging apps have experienced this technology, the most well-known being Siri by Apple, Alexa by Amazon, and Cortana by Microsoft.

Bots like Siri are used millions of times each day, answering questions and providing information. While these friendly, sometimes witty, robots may seem intelligent, they fall short when it comes to handling a more complex request. Ask a literal question like “What movies are playing near me?” and you may get a good response. Ask for something more akin to a recommendation like “What movie playing nearby would be good to see with my kids?” and you’ll quickly stump the bot. In contrast, a human personal assistant (think concierge) would easily be able to provide a high quality recommendation and then take care of every detail of organizing an afternoon out.

Context is King

What makes it so difficult for a computer to be programmed to answer more complicated questions that a personal assistant could easily handle? After all, there are oceans of data available for access on the internet! In many ways, it comes down to context, the extra information surrounding a word or idea. As human beings we often take this unspoken information for granted because we are able to draw upon years of accumulated knowledge to understand implied meanings. In normal conversation, humans string together ideas and concepts and keep them in mind as the conversation progresses. Imitating this skill is one significant hurdle faced by researchers and developers of today’s most promising artificial intelligence programs.

Machine Learning: The Key Ingredient for AI

When computers imitate human language and conversation, they appear to be acting independently. In reality, computers can only complete the tasks that they are explicitly programmed to perform. Developers code a myriad of scenarios that enable the computer to “intelligently” answer questions.

Obviously, thinking of every possible combination or question is impossible – thus the long-held desire to enable a machine to “learn” from previous queries. Known as machine learning, this field of computer science seeks to give computers the ability to accomplish new tasks without being reprogrammed. Machine learning leverages pattern recognition and the construction of algorithms that enable predictions based off of dynamic data inputs.

Viv Unveiled

Machine learning isn’t new, but applying it to create a digital assistant is no small task. Learning occurs by spotting patterns and repeating actions when triggered by key words or other stimuli. So, if the input is poor, the machine’s behavior evolves accordingly – as evidenced by Microsoft’s failed experiment with Tay.

Another attempt has quietly been in development by Viv Labs, a company led by two of the creators of Apple’s Siri. Recently revealed, the project aims to “radically simplify the world by providing an intelligent interface to everything.” The team hopes to make a significant leap forward by providing an AI platform that connects devices, services, or even things everywhere.

Rather than being a closed platform with limited intelligence on a wide range of topics, Viv is being developed as an open environment that intentionally links together different services. For instance, if you found yourself unexpectedly needing to travel, Viv might query several of your preferred providers for car rental, trains, and flights, then advise you which is closest, cheapest, and/or most quickly available.

While all the building blocks of artificial intelligence have yet to be realized, advancements are occurring at record pace. Soon enough, the commonplace bots of today could become the truly indispensable personal assistants of tomorrow.

If you like hearing about trends in technology, consider subscribing to our InspiringApps newsletter. Once a month we’ll send you the latest on mobile trends, technologies, and innovations.