In this article I will talk about how Artificial Intelligence (AI) is being implemented in many fields and industries. There are those who believe that AI is just hype and that it does not have any real-world implementations, so it is important to showcase how Artificial Intelligence is being utilized in various fields to help convince them that this is useful technology. This article will illustrate that AI is not just the latest in a long line of buzzwords but that it is positively impacting many industries. I will not be able to cover all the accomplishments involving AI in a single article, but I have chosen some of the most prominent.
Artificial Intelligence in Health Care
One field where AI has made a big impact is in health care and medical research. The best example to start with is IBM’s Watson, which can detect cancer more accurately and in a fraction of the time that doctors would take to read the patient’s records and draw their own conclusions. IBM Watson also creates a personalized treatment plan based on a patient’s medical reports. These plans are extremely accurate in comparison with what doctors would prescribe. IBM Watson’s personalized treatment plans can create the plan in just 10 minutes whereas a doctor would need on average 160 hours to create a similar plan. IBM Watson’s amazing personalized treatment plans were a point of discussion at the annual meeting of American Society of Clinical Oncology.
A study was done in India using IBM Watson treatment plans. When they were compared with plans doctors would recommend for patients, the Watson plans were found to be accurate 96% of the time for lung cancer, 93% of the time for rectal cancer, and 81% of the time for colon cancer. It is not just companies like IBM using AI in health care solutions, but Google is also not very far behind. A new algorithm developed by the Google AI team can aid in the detection of eye disease.
We live in a world where there is severe shortage of doctors who can attend to and take care of patients. Think how great it would be if we had AI based Robot Doctors (RoboDocs) like RoboCop who could attend to patients in their stead.
Do not be surprised if in the future you may not need to visit a doctor for normal diseases. An AI implant in your body (nano-technology will come into the picture in situations such as this) will record the required medical parameters for your body and send them to a system, which can recommend a course of treatment. It will repeat the same procedure again and again until your medical parameters are back to normal. With the above few examples of how AI is bringing massive changes to health care, you should be convinced that AI is making a meaningful impact.
Artificial Intelligence in Transportation
The transportation industry seems to be first industry that will have to go through the test of trusting AI on a large scale. This industry will be a huge test case of AI investments and will come to define the future of AI and how the public perceives it.
Whenever we hear AI we typically imagine two things: robots and self-driving cars. It seems that the auto industry is in awe of AI as most of the major automobile companies are looking to roll out their own self-driving cars or trucks. Many of these companies are already in the advanced stages of testing and have shown great results in their test runs.
But AI in transportation is seen with suspicion as well. Since it involves huge risks to public safety, very bold legislative decisions by governments all around the world are needed to allow self-driven cars on public roads. Another perceived outcome is the loss of jobs for human drivers without the creation of alternative jobs, which will surely invite lots of protest from many quarters of society. The human race has seen countless changes that have been resisted throughout history, but history has shown us that stopping technology use cannot be resisted for long.
Self-driving automobiles are bound to become a reality of life as companies like Google, Tesla, and Uber are planning to begin rolling out their own self-driving cars in 2018. Ford has announced that it would begin producing cars without steering wheel by the year 2021, believing that by that time self-driving cars will have become the norm and not the exception. Ford has doubled its investments in R&D in this regard.
Implementing self-driving transportation on a large scale will certainly be a challenge. Questions about whether it will decrease accident rates, relieve traffic, lower energy consumption, or how legal issues such as accidents will be handled all must be answered. According to a study conducted by Morgan Stanley, self-driving cars would save the U.S. $1.3 trillion a year by 2035 and $5.6 trillion globally on an annual basis.
In October 2016 Otto, an Uber-owned company, used a driverless transport truck travelling 120 miles from Fort Collins to Colorado Springs to successfully deliver 50,000 cans of beer. The truck was accompanied by a human passenger in case of the experiment were to go wrong. The truck itself was tuned to a maximum speed of 55 MPH. This experiment is a wakeup call for the transportation industry. In Feb 2018, Embark’s self-driving truck completes test drive from Los Angeles to Jacksonville (2400 miles). It will be interesting to see if AI drivers go beyond self-driving cars to self-driving trains, passenger airplanes, and ships.
Artificial Intelligence in Manufacturing
When it comes to large scale AI usage, it is the manufacturing industry with their industrial robots that takes the lead. By robots I do not mean what you would conventionally imagine with two hands, legs and camera eyes. These industrial robots come in many different forms and sizes. Today’s intelligent robots have the capability of assembling complex products like electronics and cars.
Industrial robots have been used since the 1960’s but what we see today are more intelligent AI-enabled robots that can work with precision and hardly need any human monitoring to complete their tasks. These robots are ever increasing productivity and accuracy on the assembly lines. Equipped with sensors and other gadgets that are connected to various computers, robots keep on tracking their progress, record required data about their work, and provide analytics to management about manufacturing statuses, defect detections, corrective measures, and reports.
Cost cutting is a primary concern for any industry, giving a competitive edge for robots in the marketplace over human workers. Robots make the entire production cycle faster and more efficient, working 24/7. For the same things to be done with human workers, a company would be required to hire humans for 3 shifts, costing huge expenditures associated with the workforce that often do not apply to robots. China has largely adopted robots in their manufacturing industries and has invested billions of dollars in industrial robotics. A study by Frost & Sullivan’s says that the global industrial robotics market is expected to increase by two-fold and could reach up to $70.26 billion by 2023, growing at a compound annual rate of 8.1 percent.
I would like to share an interesting development done by OpenAI researchers that you might think is straight out of a science fiction movie. These researchers are working on “one-shot imitation learning” in which robots learn to do tasks simply by watching humans do them. This advancement in technology could bring massive change in robotics. Here is the link where you can view this amazing feat: https://blog.openai.com/robots-that-learn/.
Artificial Intelligence in Customer Service
A major provider of jobs around the world is customer service, as every industry has a department to deal with their customers. As per the statistics of the Bureau of Labor, over 2.7 million Americans who are employed as customer service representatives with a mean salary of $35,170. These jobs are now being challenged by AI and have already seen its impact in their customer service Industry.
As of today, the primary implementation of AI in customer service or customer care is the deployment of machine learning based chat bots that are interacting with customers. Right now, these bots can sort routine customer service requests and customer grievances, but this is only the initial stage where these bots are being deployed for minimal and simple customer queries. But even with these queries the bots are doing an excellent job with a high success rate.
The adoption of bots has become easier with the availability of Chatbot platforms such as IBM Watson, Microsoft Bot Framework, LUIS, Wit.ai, Api.ai, and Chat fuel. Some of these platforms are free and open sourced, allowing even non-technical people to create bots for their businesses and deploy them as independent applications or integrate them with well-known chat applications like Facebook messenger or Slack. These chat bots are not just limited to replying to customer questions but are now becoming more advanced such as the chat bots deployed by banks that are able to perform banking transactions while interacting with the customer and respond to certain queries.
These bots self-learn as they continue to interact with customers making them more intelligent, smarter, and accurate. The biggest factor in the rise of chat bots is due to popular messengers like Facebook messenger opening their platform to integrate with third party chat platforms. This opening makes it easier for customers as they will not need to install any other pieces of software to interact with their vendor’s customer service.
A study by Gartner in 2016 revealed that nearly 44% of the respondents to their survey said they would be implementing virtual customer assistants in the next 2 to 3 years. According to a study conducted by McKinsey in 2016, 29% of customer service jobs have the potential of being automated.
Artificial intelligence in Cyber Security
There is a very tricky situation when it comes to the usage of AI in Cyber Security: a game of cat and mouse. Cyber security experts use AI to create more self-learning and adaptive threat detection. Cyber criminals also utilize AI to make their malware and viruses more sophisticated. As criminals make sneakier attacks the crime fighters will be able to detect these threats earlier and earlier.
Artificial Intelligence and Machine Learning are set to make cyber security more intelligent, allowing the prediction of threats with greater speed and accuracy before major breaches and damage are inflicted by cyber attackers. AI is bringing better solutions to the table in countering cyber-attacks.
In this new era of AI-based security many startups are developing creative cyber security solutions. Even the well-established names in the software industry are putting a major impetus on adding AI-based cyber security solutions to their existing applications.
Watson for Cyber Security is a AI-based system that uses the cognitive capabilities of Watson in providing companies greater insights and analytics in securing their enterprises. Microsoft Intelligent Security Graph is a AI Machine Learning enterprise cyber security solution that provides security to Microsoft products and offers to detect threats earlier and respond faster. According to Microsoft their Intelligent Security Graphs learns and improves every month as Microsoft analyzes over 450 billion authentications, 400 billion emails scanned for malware and phishing, and 1 billion windows devices updated.
Further proving how universities are the fountain head of many path-breaking research we have had in various fields, researchers at MIT have developed a new AI-based cyber security platform called “A12” that has the capability to predict, detect and stop 85% Cyber-attacks.
In a recent survey by The Global State of Information Security Survey, it was found that almost 23% of companies are willing to invest in and are planning to use AI-based cyber security solutions within the next year. Previously there was distrust in the capabilities of AI in providing meaningful solutions in cyber security and in general, but since the companies involved in developing AI-based cyber security solutions have demonstrated and convinced others about the benefits of investing in AI. The recent ransomware attacks around the world, the hacking of Sony data, and other incidents have forced many companies to make large investments into cyber security rather than repent later and face losses and the trust of their customers.
How Tech Giants are using Artificial Intelligence
As you may be aware Artificial Intelligence and Machine Learning are not new technologies in and of themselves, however there has been a massive surge in recent interest. Many top companies have been using it internally in many of their products or services over the years, but now the very same companies are coming up with tools, platforms and other offerings for customers to use and develop their own solutions. These companies compete to acquire other smaller AI companies so that they can have any sort of edge in the market.
In their annual report filed in June 2017 with United States Securities and Exchange Commission, Microsoft stated the following as their vision for the company: “Microsoft is a technology company whose mission is to empower every person and every organization on the planet to achieve more. We strive to create local opportunity, growth, and impact in every country around the world. Our strategy is to build best-in-class platforms and productivity services for an intelligent cloud and an intelligent edge infused with artificial intelligence (“AI”)”. The emphasized text clearly shows the new vision of Microsoft is to give priority to Artificial Intelligence. Microsoft currently has around 7,000 data scientists and engineers around the world through their own research as well as in partnerships with universities dedicated to working on this AI-infused vision.
Microsoft has built their machine learning capability over the years, quietly integrating it unobtrusively with products that many people were using. Major products such as Bing, MS Office, and Xbox games all benefit from these integrations. In 2015, Microsoft made its Azure Machine Learning platform on its Azure cloud available for public use for customers to build their own custom solutions, marking the first step towards democratizing machine learning. In 2016, Microsoft also released the Microsoft Cognitive Toolkit that you can read more about on their blog: https://blogs.microsoft.com/ai/microsoft-releases-beta-microsoft-cognitive-toolkit-deep-learning-advances/. Uber uses the Microsoft Face API, which is part of Microsoft Cognitive Artificial Intelligence services, to ensure the driver using the app matches the account on file. According to Uber it was extremely easy to integrate this API into their platform, taking just 3 weeks to perform the integration with their systems.
Microsoft also provides the Microsoft Bot platform to build, deploy, and integrate chat bots with Skype, Facebook Messenger, and other popular messaging apps.
IBM has a variety of businesses and product lines such as IBM Mainframes, Middleware, Servers, and many other enterprise tools. IBM is one company that has major stakes in AI in terms of its vision as well as investments providing Artificial Intelligence, Machine Learning, and related services under its IBM Bluemix cloud platform.
IBM Watson for Oncology, a technology capable of diagnosing cancer and creating personalized treatment plans for patients, has been in continuous developed for the last six years and is being deployed in many cancer hospitals around the world. According to IBM, in the near-future Watson for Oncology will be effective on 12 types of cancers, which currently account for 80% of the world’s cases. This technology has had billions of dollars invested in its development but has not yielded any expected financial results so far. Overall IBM Watson as a technology is not impacting AI business around the world and the costs associated with the implementation of using Watson are overrunning any profits for IBM. Despite this, IBM believes that it is bound to make good business in the future will continue to bet big on Watson.
In the USA, every year, 1.55 million active military members make the transition from military to civilian life. An IBM Watson-based engagement advisor is helping military men and women in this transition. The IBM Watson-based advisor helps them answer important questions as well as exposing them to any rules, laws, and schemes that they can take advantage of in this transition.
DBS Bank in Singapore was the first customer in Asia to deploy IBM Watson in their banking services for better client experiences. DBS bank is currently known as the world’s best digitized bank. They use various services provided by IBM Blue mix and Watson for their day-to-day operations.
IBM has also recently started testing an innovative AI model that can mimic the human brain. Using neuroscience techniques to aid in the model, IBM has dubbed it Neuromorphic computing. Working with US air force, IBM hopes to build a Neuromorphic super computer.
Under their Bluemix platform, IBM provides various services related to Machine Learning such as Watson Alchemy that is used for Natural Language processing, Watson Language translation for translating between languages, and Watson personality insights. IBM recently said they will be investing $240 million over a period of 10 years to research in Artificial intelligence in a partnership with Massachusetts Institute of Technology (MIT). Named the MIT-IBM Watson AI Lab, it will focus primarily on four areas of AI research: new algorithms, hardware, social impact, and business applications.
If there is one company that is on an acquiring-spree of AI startups around the world it is none other than Google. Google acquired eleven AI startups in 2016 alone and it is continuing this trend of buying out AI companies. Google has been quietly implementing Machine Learning into their products such as Gmail, Google Ads, YouTube, Google Play, Google Search, Google Maps, Google Assistant, Android, and Google Photos. Google’s new motto seems to be “AI First” and Google has been revamping all its products with this as their objective, making their products better and better. The biggest asset Google has is all the data collected on their search engine and the Android platforms, which make it easier for Google to improve their products using AI since data is the most important input to AI.
Google provides its AI and ML offerings through its Google cloud platform and have open sourced their core machine learning framework TensorFlow. TensorFlow has now become one of the more popular frameworks, with companies such as Intel, SAP, Snapchat, eBay, Twitter, Qualcomm, Airbnb, and Airbus using the TensorFlow framework to build their machine intelligent applications. The TensorFlow framework can be used to build voice, sound, and image recognition, video detection, and text-based applications like language detection, natural language processing, and text summarization.
In 2014, Google acquired a British AI startup DeepMind that was doing research in deep learning for a whopping $400 million. This was a hotly debated topic in the tech industry when DeepMind was used to build the AI model AlphaGo to play the game Go, as AlphaGo eventually defeated the best Go player in the world. Many could not believe that the AI model could self-learn so fast, making counter moves and defeating the human player. When this tussle of machine vs human was initially announced many were very skeptical about DeepMind’s capability to defeat the best player in the world.
Google DeepMind was implemented within Google’s data centers to manage power consumption, producing astounding results with power savings up to 40%, which by any standards is a great achievement. These power savings were found by feeding the cooling system sensor data to DeepMind, giving DeepMind the ability to predict when to regulate power in these systems to help save power.
Google has collaborated with London’s Moorfields Eye Hospital where the hospital will provide access to around one million eye scan images for DeepMind to be trained on, helping doctors to detect eye aliments more easily while treating patients. Google has also stakes in self-driving cars that they have been testing for quite a long time on Google’s campus. So far Google has tested over 700,000 miles without any fatal accidents.
Another interesting Machine Learning offering from Google is its Jobs API that will accelerate the rate of employee hiring at companies. The Jobs API empowers companies to match ideal candidates to their job openings by comparing many parameters that are generally missed by more traditional job portals. It not only makes recommendations about candidates who would be a good fit for their job role, but are also more likely to join them. For now, the Jobs API is still marked as being in alpha testing.
For many non-technical people, Amazon is best known as an online retailer website. The more tech-savvy individuals know Amazon more for their Amazon Web Services (AWS). Amazon has also jumped into the band wagon of Artificial Intelligence and Machine Learning as it seems obvious that no major tech giant wants to be left behind in AI. This may not come as surprise but the product recommendations on Amazon.com comes from its Machine Learning engine that Amazon has been working on over many years.
Over the last 20 years, Amazon has been investing much of its research in deep learning. A majority of this research has been used internally in applications such as their automatic order fulfillment centers that are automated by robots picking routes and packaging orders. Amazon’s supply chain, forecasting, and planning have also been strongly based on Machine Learning algorithms. Their drone delivery system is a good example of multiple AI technologies being used together to achieve a single goal. Amazon’s personal assistant Alexa is another example of multiple AI technologies working together, being powered by Natural Language processing implemented with a deep understanding of voice recognition algorithms.
According to Amazon, “Amazon Go” is the world’s most advanced shopping technology. Again, Amazon Go is a combination of many technologies like computer vision, deep learning, sensor fusion, and geo-fencing. Amazon Go allows customers to select products and automate their payment without interacting with any humans in the store.
In 2016, Amazon opened its Machine Learning platform for the public on its famous AWS cloud. Just like others, they too, have services for image, speech, and text recognition, chat bot development, and running third party Machine Learning algorithms on the servers using AWS. Amazon Rekognition is a service based on deep learning provided as part of the Amazon Machine Learning platform that allows users to analyze images within your applications, detecting objects, faces, and mark images as appropriate or inappropriate.
Amazon Polly is a text-to-speech service offered as part of AWS cloud that uses deep learning technologies to generate speech and sounds mimicking a human voice. Amazon Polly supports multiple languages and has support for both male and female voices. Amazon Lex is another service part of AWS cloud that allows developers to create conversation applications like chat bots and virtual assistants.
Finally, Amazon provides a dedicated Machine Learning platform on its AWS cloud where developers can create machine learning models using visual tools without having in-depth knowledge of Machine learning algorithms. Bundled with the power of AWS, developers can scale the computation power required to run deep learning algorithms without much worries.
Facebook started as social media platform but is now evolving into a tech service provider. Much like the other tech giants Facebook has its own AI research division known as FAIR (Facebook AI Research). The following is part of the vision of FAIR as mentioned on their website: “Our research covers the full spectrum of topics related to AI, and to deriving knowledge from data: theory, algorithms, applications, software infrastructure and hardware infrastructure. Long-term objectives of understanding intelligence and building intelligent machines”.
In September 2017, Facebook announced that they have developed an AI system that can react to human expressions. This system monitors 68 points on the human face for changes. This technology has the potential to greatly help in conversation applications to help gauge mood and reactions of participants.
In a rare collaboration between competitors who are each trying to carve out their own spaces in AI, Facebook and Microsoft have come together to join forces to provide an open source repository of algorithms that will help democratize AI. Called the Open Neural Network Exchange (ONNX), this repository will allow developers and companies to freely use algorithms that fit whatever their use case may be.
Hopefully this article has helped you to understand how AI is making its mark on the world, from industrial machine lines to disease detection and treatment to customer service. All the big tech players are using AI to help create more intuitive products and make better business decisions. Finally, I hope that this article has sparked enough of your curiosity to begin delving into the multitude of AI technologies out there to join in the further development of AI and advance your career to new heights.