3 Big Trends Shaping the AI Ecosystem Right Now

DZone 's Guide to

3 Big Trends Shaping the AI Ecosystem Right Now

Let's look at three trends that are shaping the AI ecosystem right now, such as exponential growth, increased accessibility, and AI integration becoming the "new normal."

· AI Zone ·
Free Resource

From ride shares to smart power grids and from healthcare to our online lives, AI is being propelled out of labs and into our daily lives. Microsoft is betting that conservation-focused AI can save our planet, while Facebook sees it as a silver bullet for rooting out harmful content. Tesla CEO Elon Musk and the late physicist Stephen Hawking both warned society of the potential for weaponized AI.

At CA, we wanted to gain insight into how the AI Ecosystem has developed over the past year. We partnered with Quid, a San Francisco-based startup, whose platform can read millions of news articles, blog posts, company profiles, and patents — and offer immediate insight by organizing that content visually. From its global dataset of 1.8 million companies, Quid classified companies that mentioned a specific focus in "Artificial Intelligence" or "Deep Learning."

"Partnering with CA on this project allowed us to cut through the hype and see where the real growth in AI and deep learning are taking place," said Quid CEO & Founder Bob Goodson.

From our analysis, three major points stood out: exponential industry growth, increased accessibility to AI technology, and adoption of AI as the way of today's business. Let's take a close look at these points.

#1 Industry Characterized by Exponential Growth and Investment

The number of companies in the AI Ecosystem has more than doubled in the past year. According to Quid, companies that referenced use of "Artificial Intelligence" or "Deep Learning" jumped from 2,809 to 6,955-a 250% increase.

Investments and the number of private investors in AI have rapidly accelerated. Over the last year, there were 1,027 investments totaling $10.5 billion, compared to 1,993 investments totaling $14.1 billion for the previous decade. Seen another way, this represents an average of $10.2 million per investor in the last year, contrasted with $707,000 per investor per year from 2007 to 2017.

The areas that have received the most overall global investment are Image and Facial Recognition, FinTech, Neural Network Tools, and Self-driving Cars. Investments in Image and Facial Recognition technologies totaled nearly $1.3 billion while FinTech received over a $1 billion, primarily from the UK and US. If we remove extremely large single company investments, the top four areas for AI investment are Neural Network Tools, FinTech, CyberSecurity and Industrial Automation/Robotics, which received around $2.5 billion combined in 2017.

While the United States remains dominant in the AI Ecosystem, China has emerged as a major player dedicating $1.1 billion to image and facial recognition technologies alone. Alibaba's recent investment in SenseTime aligns with China's vow to become a world leader in AI by 2030.

#2 Increased Accessibility of AI Technology Driving Rapid Adoption

The widespread availability of toolkits, particularly for Deep Learning applications such as voice or facial recognition, is democratizing access to sophisticated AI technology. In the past, companies only applied AI to valuable problems, because it was an expensive tool to use and often required a team of PhDs or data scientists to produce. These days, an app developer with access to cloud-based AI services-tech giants Google, Microsoft, IBM, and AWS all offer them-is sufficient.

Two notable additions to the AI Ecosystem from the past year were clusters that specifically formed around the Alexa and Watson technologies from Amazon and IBM. 155 companies were linked to Alexa and 50 companies for Watson.

Both companies released assistants in 2018. The new Watson Assistant gives developers the power to build voice activation into their own apps, and Alexa's Mobile Accessories toolkit allows users to integrate Alexa directly into wearable technologies.

#3 AI Integration Becoming the "New Normal"

AI technologies are being adopted across a wide spectrum of industries, from conducting legal due diligence to healthcare, real estate, marketing, and even agriculture. AI is becoming an integral part of doing business, and its adoption follows similar paths forged by the invention of the internet and later, mobile technologies. Soon, you won't think about building a product or solving a problem without using some (perhaps many) of these capabilities.

One reason for the rapid industry growth over the past year is that we're seeing fundamental shifts in the way companies are approaching AI. More than a quarter of the companies that began identifying as AI or Deep Learning between Q2 2017 and Q2 2018 were founded before 2013; a third was founded before 2017. This means they either changed their messaging to leverage connections to a popular industry or, more importantly, they made a decision to begin integrating AI into their products and services.

CA Strategic Research and Bias-Resistant AI

We are rapidly moving into a world where we are subject to decisions made not by people, but by algorithms. The AI landscape today is replete with examples of unintended bias, making it a minefield of ethical issues. At CA Strategic Research, we are focusing on decreasing biases in AI.

Our three-year, EU-funded ALOHA project (adaptive and secure deep learning on heterogeneous architectures) deepens the understanding of how applications running on IoT devices with growing computational power can learn from experience and react autonomously to what happens in a surrounding environment. The research aims to create tools and methods that will help AI developers and business managers lower the likelihood of creating biased AI applications. The ALOHA project has huge implications for "ethical AI" and creating software that is more transparent and "bias-resistant."

To find out more about CA's work on bias-resistant AI, download our booklet.

ai adoption, artificial intelligence, deep learning, machine learning

Published at DZone with permission of Otto Berkes , DZone MVB. See the original article here.

Opinions expressed by DZone contributors are their own.

{{ parent.title || parent.header.title}}

{{ parent.tldr }}

{{ parent.urlSource.name }}