The Data-rich Are Getting Richer – The Current State of Artificial Intelligence (Part 1)

mi-landscape-3_7

 

You probably haven’t missed that there’s much discussion about Artificial Intelligence (AI) and Machine Learning these days. In a two parts blog we take a look at what AI is, and it’s effects, then explore the state of AI in the Nordics.

We think that Artificial Intelligence will have significant impact on economies, companies and individuals during the coming years, and create a new, increasing digital divide between the “Haves” (data, software, capital, resources, talent) and the “Have-Nots”. In short, the data-rich are getting richer, and in this blog post we aim to explain why.   

Everybody uses buzzwords like AI and Big Data, but what is it really? To start with the basics, we invited Jay Solomon of Augify, an AI-company based in London and Stockholm, and sat down to talk about AI and Machine Learning.

Artificial Intelligence is about creating computers and software capable of intelligent behaviour. We use the term for machines that mimic cognitive human functions like learning, hearing and seeing. In recent years, Deep Learning has helped us to vastly improve computer vision, language, and speech understanding. Deep Learning is a set of algorithms that are inspired by the workings of the brain and how we learn. What it really means? A new world of self-taught machines. Instead of programming computers to perform particular tasks, as we are used to, we program the computer to know how to learn. This is what we call Machine Learning.

At the core of Deep Learning are Neural Networks, which basically is a computational approach that also mirrors the brain. AI and Neural Networks have become hugely powerful thanks to two advances in recent years. 1) A better understanding of how to fine-tune networks as they learn, thanks in part to 
much faster computers, and 2) The availability of massive databases (“Big Data”) to train the networks. Artificial Intelligence feeds on data, and the more data the systems can digest, the more they learn and improve.

To make it simple, the raw combination of data and computing power makes AI stronger every day. This has inspired the popular doomsday future scenario when Terminator-like machines run by “Skynet” takes over the planet, when we have lost control of our creations.    

If you’re not reading any further than this, let’s summarise the overview above with five basics concepts that will at least make you look smarter at any dinner conversation.

  • Artificial intelligence (when machines mimic human, intelligent behavior)
  • Machine Learning (programming machines to learn how to learn)
  • Deep Learning (networks and algorithms inspired by the functions of the brain)
  • Neural Networks (interconnected networks, also based on the human brain)
  • Algorithms (set of rules that defines computational operations)   

These five concepts are all related, as Artificial Intelligence is the general term at the high level for machine intelligence, Machine Learning is about acquiring this intelligence, i.e. self-taught, autonomous machines, Deep Learning and Neural Networks are the algorithms and architecture inspired by the human brain used for Machine Learning, and the algorithms themselves, at the most granular operational computing level, is the code, the software.

Another way to see it then, is that Artificial Intelligence is software. This “AI-software” is now used to empower other, existing software to become more intelligent. A typical example is when a social platform like Facebook, with enormous amounts of data, deploy digital assistants to make more use of this data for its consumers on the platform. For examples, see our previous blog The Meaning of Bots. That AI is just data and software, although very powerful and paradigm-shifting, has inspired the phrase “Software is eating the world”, made popular by Marc Andreesen.

Our world as we know it is already
 running on Artificial Intelligence. Siri manages our calendars, 
Facebook suggests our friends, computers manage our pension funds, cars can now park themselves and air traffic control is almost fully automated. AI is becoming embedded 
everywhere. AI is used in many and various areas such as speech
 understanding, machine
 translation, computer
 vision, handwriting
 recognition, face
 recognition, natural language understanding, automated copywriting, auto generated
 ad campaigns, and as we have seen, intelligent
 (and not so intelligent) bots.

People are also becoming aware of the negative side effects of algorithms that feed us information in “filter bubbles” according to our automated profiles on social networks and in search engines, that cannot even distinguish between real and fake news. This has in turn started a discussing about how good these algorithms actually are, prompting views, for example, that Facebook really suck at machine learning. But make no mistake, the global digital giants invests aggressively in the AI race to gain advantages.        

As AI is now everywhere, has given rise to another idea – that AI is the “new electricity”. This concept is being promoted by Jack Ma, the founder of Alibaba, and also explored by Kevin Kelly in his book The Inevitable, about the tech forces that will shape our future. One of these mega trends is “Cognifying”, or the evolution of adding cognitive, human-like skills to software.

As Kevin Kelly puts it, “The AI on the horizon looks more like Amazon Web Services—cheap, reliable, industrial-grade digital smartness running behind everything, and almost invisible except when it blinks off. This common utility will serve you as much IQ as you want but no more than you need. Like all utilities, AI will be supremely boring, even as it transforms the Internet, the global economy, and civilization. It will enliven inert objects, much as electricity did more than a century ago. Everything that we formerly electrified we will now cognitize. This new utilitarian AI will also augment us individually as people (deepening our memory, speeding our recognition) and collectively as a species. There is almost nothing we can think of that cannot be made new, different, or interesting by infusing it with some extra IQ. In fact, the business plans of the next 10,000 startups are easy to forecast: Take X and add AI. This is a big deal, and now it’s here.”

Take a toothbrush and add a motor = Electric Toothbrush. Take a house and add AI = the Smart Home. Take enterprise software, like HR tools, and add AI = the Robot Recruiter. For example, Mya is an AI-tool designed to automate much of the recruitment process.

And sure enough, Google Cloud has already announced new machine learning features for enterprise use, offering “AI on tap” letting you rent your own machine learning computer, complete with APIs for tasks like translations, job matching and analysis. Google’s CEO Sundar Pichai recently got much attention for stating that they are moving from a “mobile-first” company, to AI first, adding more intelligent functionality to all their core products.

However, Andrew Ng, one of the world’s leading AI experts, professor at Stanford and Chief Scientist at Baidu Research in Silicon Valley, suggests what AI can do for companies right now. “Despite AI’s breadth of impact”, says Andrew Ng, “the types of it being deployed are still extremely limited. Almost all of AI’s recent progress is through one type, in which some input data (A) is used to quickly generate some simple response (B)”. For example: input A (picture), response B (“Are there human faces?”), application (photo tagging), or input A (car camera), response B (“Position of other cars?”), application (self-driving car).  

The field of Artificial Intelligence is some 60 years old, but it’s not until now that it’s really happening on a broader scale and enter our lives. May we ask why? In a recent podcast from Silicon Valley investors Andreessen Horowitz (When Humanity meets AI), Stanford associate professor Fei-Fei Li (another AI expert now hired by Google) explains that we now are at “a historical moment in AI”. Three forces combined trigger the switch. First, AI technology (the deep learning and neural networks mentioned above) has come of age, second, Big Data feeding the technology is now available, and third, computing hardware (processors and deep learning chips) have developed fast in recent years. One of AI’s next big frontiers will be chips.

So, what’s going on in the AI industry? To answer that question, we have to look at what’s going on primarily in USA. Recently, Sam DeBrule (curator of the weekly Machinelearnings newsletter) posted a “Humans non-technical guide” to the state of artificial intelligence in the US. This summary of the Machine Intelligence Landscape (pictured above) included some 320 AI & ML companies, active in fields like Enterprise Intelligence (visual, sensor), Enterprise Functions (sales, recruitment), Autonomous Systems (navigation, robotics), Agents (personal, professional), Industries (education, investments), Healthcare (patients, images) and Technology Stack (natural language, data science). The list also featured some 80 influential people, the most famous probably being Elon Musk who have launched the OpenAI initiative, and about 30 AI news sources. In addition, the US administration has launched an ambitious report and a strategic plan to support AI.

But the most intensive race for AI is going on at the world’s largest software companies. The world’s top-3 largest companies by market capitalisation today are Apple, Alphabet (Google) and Microsoft. Facebook and Amazon are on the top-10 list (and somedays top-5), and their combined value is currently close to 2,500 billion dollars. And they have huge coffers ready with cash to spend. These companies have built their value on global digital networks with products and services used daily by billions of people on the planet. They are now rapidly absorbing startups, resources and experts in AI and ML to stay ahead. Much of the research and practical application of AI is now happening within these corporations, further establishing their already dominant positions. They have the ability to use AI as an extremely powerful leverage on their enormous amounts of data, and as the the data volumes to grow, as it will, the systems become even stronger. AI is an catalyst for an escalating edge that is leaving other companies, and nations, behind in the race.

To conclude, we think that the developments within AI and ML outlined above will have far-reaching effects on society, industry and for consumers. At least four, long term effects are pretty clear.           

  • Companies that already have access to massive data will become even more powerful as AI leverages the data (Google, Apple, Facebook and others)
  • We’ll see more AI-startups, enabled through access to increased computing power, third-party data, cheap cloud storage, and open source AI software
  • Commercialization of AI for consumers will offer an exploding supply of new intelligent services and products, as AI adds to existing software
  • Nations with a political agenda (research, investments, infrastructure, legislation, education, etc) for AI will get an advantage

However, we are not sure about how slow-moving established companies will reap the benefits of AI, or how they will keep up in the race. Furthermore, countries that are already lagging behind will see a growing gap to the most advanced AI-nations.

The nature of AI in combination of Big Data, drives an ever increasing advantage for companies and countries already with advanced AI-technology, research, data and capital, like the US and China. The dynamics of AI and Big Data moves us towards a new global digital world order, further shifting the power balance to the already dominating tech regions like Silicon Valley and Shanghai.         

So in a potential new or reinforced digital world-order, where the already data-rich get richer, what about Europe? And will the Nordics keep up and prosper with new opportunities in AI? We now turn to our own region and wonder what the accelerating AI-development means for both companies and investments here. In our next blog, “The Current State of AI, Part 2 – Is There Intelligence in the Nordics?” we explore what’s happening, and maybe not happening. 

Image courtesy of Shivon Zilis and James Cham, designed by Heidi Skinner