Your voice assistant chaperones

Welcome to the Party, here are your voice assistant chaperones.

It’s hard to pinpoint the exact path that Artificial Intelligence Voice Recognition will take.  But with companies such as Google, Amazon, and Microsoft making huge strides in voice assistants and the late Stephen Hawking and Elon Musk calling for AI to be regulated and monitored, it is definitely a hotly debated topic.

The ultimate goal of Artificial Intelligence (AI) is to be omnipresent in someone’s life. AI should be there to provide what you need in order to do what you want to do. Today, hardware is the means to which AI presents itself into our lives. Siri, Alexa, and Google Assistant are the consumer-facing portals to which we access these algorithms. Furthermore, Apple, Amazon, and Google are racing to add their digital personal assistants into smartphones, desktops, laptops, tablets, and an ever-increasing variety of smart speakers and speaker-screens, cars, TVs, home security systems, thermostats, light bulbs, kitchen appliances and more.

Apple 

There is little argument that Apple is winning the hardware race. With over 1.3B active devices, as of January, 2018, and a head start in the smart assistant race it would be logical to think Apple is also winning the AI battle. The Cupertino based company, however, didn’t take advantage of the early lead they had with launching Siri in 2011.

The combination of the company’s preference towards security and privacy plus the amount of secrecy surrounding upcoming product launches isn’t the type of environment that top deep learning experts want to operate in. Those who want to solve the problem of general AI also want to have their work published and acknowledged by the research community.

Apple Cash Resources

Apple is now playing catch-up against other tech companies in the race to implement AI. This is the toughest battle it’s been in since the 1980s race with Microsoft and IBM for personal computer dominance. Apple has the largest cash pile on the planet.  Apple can easily use much of its resources to stay competitive in a market that could contribute $15.7 trillion to the global economy in 2030.

We can already see this actualizing. In late 2016, Apple hired Ruslan Salakhutdinov, leading AI researcher at Carnegie Mellon, to lead its in house research. In July of 2017, Apple began bending its secrecy policy when it launched a machine learning (ML) “journal” (here).  In recent months they poached the top AI executive from Google. A huge win.

Apple Privacy Approach

Most of Apple competitors utilize huge cloud farms to provide the computational power needed to achieve world-class learning.  However, Apple has stayed steady in its on-device, privacy-focused approach.  We do not know if this on-device preference for computation is able to achieve the type of intelligence we expect from our AI assistants of tomorrow. However, their in-house mobile silicon architecture is at least a year ahead of the rest of the industry.  Experts seem to believe that is the future.

Apple’s iPhone X, with custom A11 Bionic processor, already has a neural engine processing system built in that takes care of image analysis and voice recognition. And with the launch of Core ML, Apple has given their large developer community a standard to integrate trained machine learning models into apps.

Deep Dojo’s Otto Schnurr states “with Core ML, Apple has managed to achieve an equivalent of PDF for machine learning … As a technical achievement, it’s stunning.”

Core ML is optimized for on-device performance, which minimizes memory footprint and power consumption. Running strictly on the device ensures the privacy of user data and guarantees that your app remains functional and responsive when a network connection is unavailable.

From the smart assistant that was first to the smart assistant that is winning the frequency of use battle (37% of survey respondents utilized Alexa on a daily basis).

Amazon hit the tech-world at large with an unseen right hook when it announced its Echo device with built-in Alexa voice-assistant in November of 2014. Amazon, like Apple, expects all internal work to feed into products and are tight-lipped about their research projects.

Amazon Acquisition Strategy

In 2014, Srikanth Thirumalai brought one of Amazon’s infamous six-pagers to Jeff Bezos as a proposal to incorporate machine learning into every aspect of the company’s business. Due to Amazon’s lack of talent in-house, it used its deep pockets to buy companies with expertise. “In the early days of Alexa, we bought many companies,” said David Limp, Amazon’s VP of Devices and Services. It snapped up Yap, in September 2011, a speech-to-text company with expertise in translating the spoken word into written language. Then in January 2012, Amazon bought Evi, a Cambridge, UK, AI company whose software could respond to spoken requests like Siri does. And in January 2013, it bought Ivona, a Polish company specializing in text-to-speech, which provided technology that enabled Echo to talk.

 

One of the biggest reasons for Alexa’s success has been Amazon’s ability to attract and embrace third-party developers creating their own voice-technology mini-applications (or “skills”) to run on the Echo itself.

Amazon lack of Mobile Products

Until Echo, Amazon didn’t have a track record of being able to ship mobile products that people love (minus the kindle). Their attempts with the Fire Phone and Tablet showed a lack of design chops.  Even Alexa has really rough edges in the sense that the skills-based ecosystem means the system is very command heavy. No human conversation to find a wine pairing for dinner starts with “Alexa, open {skill}.” You’ve got to navigate conversations perfectly for everything to be magical.

Amazon Web Services

But it’s early entrance and success in the intelligent assistant market has created a lot of energy within the company. Out of this has come the grand vision of how Amazon Web Services (AWS) can become the center of ML activity throughout all of the techdom.

This was inevitable for AWS. It began as a way to empower anyone to have the computing resources of any large tech corporation. We can see the success of Alexa in services Amazon now offers like Polly, a text-to-speech component, or Lex, a natural language processing engine.

Much of the internet running on AWS today.  It’s easy to say that Amazon is well positioned to be the platform of choice for anyone using their servers to build the vital ML services.  Further, it is easy to say that making the switch to competing services from Google or Microsoft can be extremely difficult.

“Our vision is to be Earth’s most customer-centric company; to build a place where people can come to find and discover anything they might want to buy online.” – Amazon Mission Statement

This early success in adapting ML as a service is a part of what makes Amazon well positioned to bring AI to the masses. Amazon’s beginnings as a book store and then focus on e-commerce, give it unmatched distribution powers.

There’s so much you can do in the Amazon ecosystem. Place an order on Alexa which a drone can pick up from a local Whole Foods (acquisition for $13.7B in 2017).  On another day, you can order books via voice and have it delivered within 2-hours.  You never leave home.  Listening to music or watch a movie on Amazon Prime. You never left the Amazon ecosystem. That’s possible today.

Google

At its core, artificial intelligence is about creating algorithms and optimizing them with data. Google has more data. And Google is the creator of one of the most famous and often taken-for-granted algorithms on the planet, Search. With a dioptic mission statement to Amazon, Google is striving to ”organize the world’s information and make it universally accessible and useful.”

The hands-down leader in the AI race today is Alphabet (Google’s parent company).

Google Search Yields Data

Even strictly by the numbers that count—talent, computing power, and data—Google is in the lead. It can afford the smartest people and because of the massive revenue generated by their Search business, can afford a large variety of projects, from drones to cars to smart software. Data scientists are called scientists for a reason. They dedicate their whole life to the potential of making a breakthrough in AI. These type of people rarely leave.

Technology is now on the cusp of taking us into a magical age, in which ML can prevent blindness, translate any language with expert skill, or even save endangered species from extinction.” Alphabet Chairman Eric Schmidt, Jan, 2018

Google’s founders were early devotees of machine learning (ML) and always saw it as a competitive edge. Machine learning is beginning to help us solve problems today that we simply couldn’t solve on our own.

Google Emphasis on AI

Google’s emphasis on AI can increasingly be seen throughout its products. Gmail suggests rapid replies to emails. Google Photos can create animations, suggest photo filters, or distinguish cherished moments to make your image library magical.

That products, however, don’t scratch the surface of the mountain that is Google’s AI prowess. Google Brain, one of their research groups, is one of the best at applying Machine Learning. YouTube viewing experience has been tuned by the studies behind Google Brain. And then there’s DeepMind, the British-based research group, which may not ever generate actual revenue for Alphabet.  But it has helped its parent save money by increasing the energy efficiency of its global data centers. (Its Go experiment was a public-relations coup).

Since Sundar Pichai took over the CEO role at Google in 2015, Alphabet has spent $30 billion building data infrastructure alone. The infrastructure likely powers Google Assistant as well as its cloud computing division and AI-backed consumer hardware lineup.

 

“Computing is moving from mobile-first to AI-first, with more universal ambient and intelligent computing that you can interact with naturally, all made smarter by the progress we are making with machine learning.” Sundar Pichai

While Google is ahead of the pack right now, the AI race is still anyone’s game. But Google isn’t taking their foot off the gas pedal. Prior to this year’s Google I/O, their developer-focused conference, they launched a rebranded homepage for Google AI.  The revised homepage highlighted consumer-focused AI products.  As well they have recently-published research in topics like health and astronomy.  

 

Voice Assistants as Home and Business Tools

In less than 5 years, voice devices and personal digital assistants have gone from a novelty technology to necessary home and business tools.  AI research continues to accelerate and the global AI race ensures massive investment. However, despite concerns about an AI Revolution, we will also see a high human component. As the volume of voice data continues to multiply, data-tagging and training needs will grow.  Easy to use, Human-in-the-Loop interfaces will become more critical. Greater systematic adoption among homes and business will continue to force innovation. The only question is who will be the big winners?

We invite you to meet Hoover, our Intelligent Assistant to learn more about swivl, how we can help you implement AI to increase competitiveness while keeping Humans-in-the-Loop.

 Join our newsletter to keep up with all the latest trends on Artificial Intelligence  powered Intelligent Agents.