Artificial intelligence (D200)

From New Media Business Blog

Jump to: navigation, search



World Economic Forum - What are AI and big data? (2016) [1]

Artificial intelligence (AI) is a form of intelligence which is carried out via machines or software. It can also be described as a type of technology or the academic field of intelligent machine studies. By definition, artificial intelligence is “a sub-field of computer science. Its goal is to enable the development of computers that are able to do things normally done by people - in particular, things associated with people acting intelligently”. [1] The term “AI” was coined in 1956 by John McCarthy who is credited as one of the founders of the discipline, along with Marvin Minsky, Allen Newell, Arthur Samuel, and Herbert Simon.

  • General AI vs. Narrow AI: Artificial intelligence is best conceptualized in two categories: general and specific - which pertain to the type of tasks the AI is able to achieve. All AI systems we encounter are currently considered narrow AI as they are only able to apply their intelligence to a limited range of problems. A popular example of a narrow AI is Apple’s Siri which is able to answer questions with the Internet as its database. IBM’s supercomputer, Watson, is also considered narrow AI as it is only able to perform as a “question answering machine” by applying cognitive computing, machine learning, and natural language processing [2]. By contrast, general AI is a “thinking machine” that is comparable to a human mind and able to act with limited computational resources in complex situations [3]. Artificial general intelligence (AGI) is ultimately the goal of AI though it has not yet been achieved.
  • Weak AI vs. Strong AI: Another distinction between AI that we can draw is whether or not it is weak or strong - explaining the capacity the AI has to mimic real human consciousness. As weak AI is only able to act as a mechanism to to emulate what humans do (i.e. play chess, drive cars, recognize speech), even the steam engine was considered a form of weak AI - albeit a very simple version [4]. On the other hand, strong AI is a the idea of a system that has consciousness like a human and like general AI, has not been created as of now.

AI Tests

Considering the formal definition of AI, most if not all programs could be considered as a form of artificial intelligence. Some researchers have developed tests to create more operational definitions for AI.

The Turing Test

Turing Test model

The Turing Test is perhaps the most recognizable part of AI history as it was the first formal test of machine intelligence. It is synonymous with the $100,000 Loebner prize which started in 1990 to judge and award human-like chatterbots. The test comprises of a human judge who communicates with a human contestant using the AI program’s methods. If the judge could be fooled into thinking he or she was really talking to a human in a freeform conversation for 30 minutes while interpreting audio-visual output, then the program would be deemed to have human intelligence. Although frequently influential and cited, the Turing Test is widely criticized because of its “black box” approach because it evaluated AI based on whether or not it behaves like human intelligence regardless of how it works [5]. Though small prizes are given to the best-performing AI programs each year, Loebner’s grand prize of $100,000 has still yet to be awarded.

The Coffee Test

The Coffee Test is one of well known scientist, Ben Goertzel’s, tests for artificial general intelligence. This test is less well known but provides a much more difficult criteria to evaluate AI. As stated by Goertzel, AGI should be able to “go into an average American house and figure out how to make coffee, including identifying the coffee machine, figuring out what the buttons do, finding the coffee in the cabinet, etc. If a robot could do that, perhaps we should consider it to have general intelligence”. This is based off of Apple Co-Founder, Steve Wozniak’s opinion of true artificial intelligence:

“These robots will kind of do one thing well, but we never will see a robot that makes a cup of coffee, never. I don't believe we will ever see it. Think of the steps that a human being has to do to make a cup of coffee and you have covered basically 10, 20 years of your lifetime just to learn it. So for a computer to do it the same way, it has to go through the same learning, walking to a house using some kind of optical with a vision system, stepping around and opening the door properly, going down the wrong way, going back, finding the kitchen, detecting what might be a coffee machine. You can't program these things, you have to learn it, and you have to watch how other people make coffee. ... This is a kind of logic that the human brain does just to make a cup of coffee. We will never ever have artificial intelligence. Your pet, for example, your pet is smarter than any computer”[6].

The Robot College Student Test

A step above the Coffee Test is the Robot College Student Test also developed by Goertzel. This test states that a successful AGI should be able to pass university being judged the same way a human student would [7].

The Employment Test

The most complex of the AGI tests is the Employment Test which was proposed by one of AI’s founders, Nils Nilsson. The Employment Test revolves around the idea that AGI should be able to do activities and tasks such as jobs which humans are employed in. Therefore, human-level intelligence should “at least have the potential [to completely automate] economically important jobs” [8]. This test would only be complete with a complete list of economically important jobs and an exam for each one. Many critics argue that this test is unfair because even humans would be unable to pass all the exams while a successful AGI would have an advantage because it would not be constrained by any variables such as limited memory, thinking speed, and sleep.

Types of AI

The seven types of Artificial Intelligence. [9]

There are seven main branches of AI that are active today: machine learning, natural language processing, expert systems, vision, speech, planning, and robotics - though the following three are the most commonly discussed in current artificial intelligence.

  • Machine Learning (ML): one of the most utilized forms of artificial intelligence, known as the "core driver of AI"[10]. It allows computers to learn by themselves without set programming to tell it what to do. Some examples that use machine learning include image analysis in Google Photos and Facebook.
  • Natural Language Processing (NLP): the ability to recognize language, its usage in various scenarios, and finding patterns in datasets. Today, AI can be used to find out customer sentiments regarding brands by analyzing unstructured data, such as Twitter and Facebook posts [11].
  • Expert Systems: system that has the capability to have deep, in-depth knowledge on a specific task or topic. The term, expertsystem, "is reserved for programs who's knowledge base contains the knowledge used by human experts, in contrast to knowledge gathered from textbooks or non-experts" [12]
  • General Benefits and Risks

    A summary of the benefits and risks of AI can be found in the table below:

Benefit Risk
Can replace human jobs in unfavorable environments High costs in creation and maintenance
Can rationalize without hindrance from emotions Multiple ethical issues about AI use (discussed in detail in the #ETHICAL CONSIDERATIONS section
Can utilize and carry out repetitive or time consuming tasks May potentially replace humans in certain jobs
Can quickly gather, analyze and derive conclusions from large sets of data There is a fear that humans will become reliant on AI for intelligence and diminish the need for humans to stop learning
AI is not constrained to human needs such as sleep, food, and breaks AI currently lack creative minds and human intuition
Can create new jobs that were not available before AI may become uncontrollable and develop a mind of its own

However, these are general concerns of artificial intelligence and we will go into further detail in each sub-analysis below.

History of AI

Caucasian Eagle [13]
Frankenstein [14]
Eugene Goostman Chatterbot [15]

AI may sound like a new technology but the concept existed ever since Greek myths were created. Below is a non-exhaustive table of the history of artificial intelligence to draw a basic understanding of the span and influence of automation and later, AI. It includes events of when AI was first hypothesized and how some of the technology slowly started to develop, although, the term, "AI," is a more recently coined term where true advancements were made in the past few decades.

Year Event
Antiquity Greek myths of Hephaestus and the creation of automatons: Talos, Pandora, and the Caucasian eagle
~900 BC Yan Shi and King Mu of Zhou’s mechanical men
~350 BC Aristotle and syllogism, formal and mechanical thought
1206 Al-Jazari’s programmable orchestra
1672 Gotfried Leibniz’s improved “Stepped Reckoner” to include multiplication and division as well as the binary numeral system and universal calculus of reasoning
1726 Jonathan Swift’s “Gulliver’s Travels” and “the Engine” which was "a Project for improving speculative Knowledge by practical and mechanical Operations"
1818 Mary Shelly’s “Frankenstein” which began to consider the ethics behind creating sentient beings
1822 Charles Babbage and Ada Lovelace’s programmable mechanical calculating machines
1944 Game Theory is introduced by John von Neumann and Oskar Morgenstern
1950 Alan Turing and the introduction of the Turing Test to measure machine intelligence
1952 IBM’s Arthur Samuel wrote the first game-playing program for checkers
1956 The first Dartmouth Conference where the field of AI research was created
1969 Shakey the Robot demonstrates animal locomotion, perception, and problem solving
1990 Rodney Brook’s “Elephants Don’t Play Chess” revitalizes the bottom-up approach to AI
1997 IBM’s Deep Blue defeats world chess champion, Garry Kasparov, though highly criticized
2002 iRobot’s Roomba introduces a new era of autonomous robots
2008 Apple’s iPhone has speech recognition (Siri)
2011 IBM Watson wins Jeopardy against two of the best contestants of all time
2014 The Eugene Goostman chatterbot raises question on machine intelligence


Marques Brownlee - Google Assistant vs Siri! (2016) [16]

There are number of artificial intelligence mechanisms already available to the general public today. A few key examples are explained below.

Intelligent Personal Assistants

Intelligent Personal Assistants (IPAs) utilizes user input, location and contextual awareness, and access to online information in order to make recommendations or perform actions using natural language. While traditional queries run through browser-centric applications, IPA utilizes components of artificial intelligence software (ex. Natural language processing and speech recognition) to complete queries [1]. Common examples of IPAs can be found integrated on smartphones or smart watches. IPAs may have the same general purpose of convenience and aiding people in their daily lives, but their personality or communication tactics can contribute to the brand or device experience. For example, in a test situation where Google’s Google Assistant and Apple’s Siri are compared to another, Google Assistant tends to be more conversational while Siri provided statistics on the phone screen instead, as demonstrated in the YouTube video to the left.


While the most common example of IPAs comes in the form of smartphones, smart homes are moving into the consumer space at a rapid rate. As companies forward the belief that for IPAs to truly bring value to the consumer, it must be universal and integrated in a variety of products, home automation is the next target. In recent news, Google Home was released on November 4th, which utilizes Google Assistant (same as the smartphone) as its AI interface. Utilizing your voice as its command, smart homes are expected to make daily tasks easier. It simplifies everyday queries, such as making dinner reservations or translating words. Depending on its integration with one’s home, it an even control lighting or temperature in certain rooms using your voice [2].

Benefits and Risks

Simply put, IPAs are common for their ease of use and quick responses. They are especially convenient for hands-free situations and can interact with smart appliances (ex. Refrigerators to remind you to buy products that are running low). However, there are a series of privacy issues set with the recording and storing of millions of voice clips. There is a loss of human features in interacting with a faceless and zero touch interface, along with the limitations of artificial intelligence as it grows in use. Finally, its integration into the internet of things (IoT) realm may also be limited as the use of IPAs is still in its growth stage that requires the commitment of consumers as much as it does businesses.

Impact on Consumers and Businesses

IPAs creates a new relationship between humans and technology in daily communication. We do not only use it as a tool to accomplish a task, but we start interacting with it on a more personal level. It provides data and a pathway for scientists and businesses to move from “computing to understanding” in why and how humans act/react the way they do. Finally, from technological standpoint, it increases integration and accessibility across various technology platforms. It brings humanity closer to the “science-fiction type of of living”, where technology is incorporated to every aspect of life.

Implications of Societal Integration

IPAs creates a new relationship between humans and technology in daily communication. We do not only use it as a tool to accomplish a task, but we start interacting with it on a more personal level. It provides data and a pathway for scientists and businesses to move from “computing to understanding” in why and how humans act/react the way they do. Finally, from technological standpoint, it increases integration and accessibility across various technology platforms. It brings humanity closer to the “science-fiction type of of living”, where technology is incorporated to every aspect of life [3]. If our searches and voices were to be leaked to individuals with malicious intent, the consequences could be disastrous. In events where our voice and personal information is used as a security measure (ex. Calling a credit card company to verify purchases using your voice and personal information), this scenario may change the way we view security and privacy.

Future Predictions

PAs are slowly being integrated into both personal and professional lifestyles, most commonly through wearable technologies. Growing rapidly in Internet services, the wearables market is predicted to grow to “485 million annual device shipments by 2018” [4]. As IPAs grow in use by consumers and businesses, artificial intelligence is learning more about human behaviour and nuances of how we act according through our voices.

Self-Driving Cars

The concept of autonomous started decades ago, in which vehicles are capable of detecting obstacles and objects in its environment and move appropriately without human input [5]. Self-driving cars utilize a combination of the following:

  • Algorithms (or rules) that tell the car, “if” this happens, then “do” this
  • LiDAR systems combined with GPS systems to detect distance between objects
  • Deep learning artificial intelligence for machines to learn how to drive itself. More notably, Tesla uses fleet learning in which vehicles will send data to the central database of how cars should react in certain situations so the entire fleet will learn as well [6]


Tesla's Autopilot [7]

While Google and Tesla makes the headlines in the realm of autonomous cars, a less-known vehicle that utilizes artificial intelligence is NVIDIA’s BB8. Begining with navigating parking lots and traffic cones, similar to how a regular teen first learns to drive, BB8 graduated to navigating roads and tricky corners through hours of experience. Although it was trained in driving in California, it was capable of navigating roads in New Jersey with no issues despite varying driving conditions. Without the use of algorithmic rules, BB8 is able to travel effectively on roads where there are no marked lanes due to its machine learning [8].

Benefits and Risks

Self-driving cars can offer potentially safer driving as it can navigate through situations with limited visibility. It can offer better utilization of space and reduce the amount of human error in overall driving. For example, if all cars on the road were self-driving cars equipped with artificial intelligence, lanes would not need be as wide and city infrastructure can be altered to save money and space. Urban planners and futurists have expressed their hopes in reinventing parking lots or roads into social areas such as parks. Carlo Ratti from the MIT Senseable City Lab predicts that “vehicle automation will require 80 percent fewer cars on any given highway… [meaning] shorter travel times, less congestion, and a smaller environmental impact” [9]. ”On the other hand, the lack of human interpretation can pose issues with human drivers that can be aggressive compared to a self-driving car, which tends to be more cautious. By having technologically integrated cars, hacking can pose a security risk for stealing cars, intentional accidents, and safety hazards in device malfunctioning.

Impact on Consumers and Businesses

A self-driving cars are becoming a greater topic of conversation, car companies are shifting gears towards developing self-driving capabilities. Companies may slowly shift their business models to align with this newer method of transportation. Self-driving transportation could also create new opportunities for companies and jobs that did not exist before. For example, popular transportation company Uber has already pushed out self-driving cars as a pick-up option in numerous cities (most recently in Pittsburgh) [10]. In order to entice users to try out autonomous vehicles, passengers that chose self-driving cars had their trip free of charge in certain cities [11]. Finally, the increased convenience of autonomous transportation could be visible for individuals that are not capable of driving on their own. Individuals with disabilities or vision impairments, for example, could benefit from autonomous transportation.

Implications of Societal Integration

While Tesla and large transportation companies are directing their efforts towards autonomous driving, there are a number of societal implications that must be taken into account. The most obvious includes the ethical implications of how computers cannot ration like humans as there is a lack of human interpretation. If an autonomous car were to get into an accident, would the passenger or the car manufacturer be at fault? Alternatively, if an accident could not be avoided and it was either the individual outside of the car or inside of the car were to be at risk of fatality, who should the car choose to save? The question of morals and ethics brings self-driving cars into a gray area without a clear black and white answer. Recently, Mercedes-Benz announced that their autonomous cars would be programed to save the passengers in the car as their first priority [12]. While this point can be justified from a corporate car manufacturer perspective that values their customers, morality continues to be questioned.

Furthermore, as autonomous cars gains more traction and buzz around the world, the need for government intervention and regulation comes into play. As this technology is relatively newer and has yet been clearly identified, the government will come to play an important role in determining what restrictions and liabilities are required of either car manufacturers or customers.

Future Predictions

The Self-driving Vehicle Revolution. [13]

In an article by McKinsey and Company (2015), it outlines the predicted growth of autonomous vehicles in terms of consumer adoption in the figure below. Currently, autonomous vehicles are used in industrial fleets, altering business models (especially true for transportation companies). Following the first era of development, consumers will begin their slow adoption of self-driving vehicles in the latter years of the 2020’s, in which post-sales service may change significantly,, altering supply chain models and insurance policies makers. Finally, autonomous vehicles are predicted to become “the primary means of transport” by the end of 2050, in which the benefits of transportation time savings, reduction of parking space, fewer vehicle crashes and coupled with cost savings, and acceleration of robotics development will be encountered [14]. This presents a more optimistic view of self-driving cars, but it may actually become a reality sooner than we may think.

Cyber Security

With the recent account of the Distributed Denial of Service (DDoS) attack on Dyn servers on October 21, 2016 (servers are home to major websites such as PayPal, Twitter, and Spotify), cyber security is becoming an increasingly important factor for both businesses and consumers alike. The DDoS attacked not only caused the outage of several networks, but also the vulnerability of information breach. The attack involved “10s of millions of IP addresses” that sent traffic to the Dyn servers [15]. According to a study by PriceWaterCoopersHouse (PwC) in 2015, 76% of US executives are more concerned about cybersecurity threats, up from 59% the year before. Approximately 79% of survey participants detected some form of cyber security incident in the past year [16].

As a result of our increased reliance on technology and networks, some are leaning towards the use of artificial intelligence to combat malicious intent and action on a technological front. According to a report by the White House released in October 2016, the government is optimistic in using artificial intelligence as a means of cybersecurity at a lower cost and increase the agility of combating attacks [17].


AI^2 Process - Adapted from Video[18]

Researchers from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) have partnered with start-up PatternEx to develop an AI system named AI^2 that can defend against cyber security attacks with greater accuracy than current methods. According to the team, the AI system can “detect 85 percent of attacks by reviewing data from more than 3.6 billion lines of log files each day and informs anything suspicious,” which is more efficient that anything that can be manually done by humans [19]. This specific system, however, uses a combination of human input and artificial intelligence to complete its task, as shown in the figure below. The AI component analyzes the data found in logs throughout the day and presents its findings to human analysts for feedback. The AI takes this into account and continues to analyze new logs, creating a cycle of ongoing feedback. In other words, this machine-learning AI utilizes data in order to become more accurate in its determination of threats [20].

Benefits and Risks

AI is capable of continuously learning and redefining information to keep it relevant and in line with new technological threats that appear every day. In particular, for AI^2, it is noted to be “3 times better than any other similar automated cyber attack detection system” in detecting threats and attacks, utilizing analyst time effectively in narrowing down the scope of logs that need to be analyzed each day [21]. On the other hand, the need for human feedback may reduce the system's efficiency in determining threats. In terms of using AI for cyber security, AI needs data and learning for maximum efficiency, which means its capabilities are limited to what it knows and is able to decipher.

Impact on Consumers and Businesses

Stronger cyber security can aid in helping to protect valuable assets, especially with the increasing reliance of the internet and technology in everyday transactions. AI is meant to assist humans against these threats, rather than replace jobs as many people fear. Businesses are placing greater emphasis on cybersecurity, which may include using AI to level the playing field.

Implications of Societal Integration

Besides the ignorance and apathy towards personal privacy and security from a consumer standpoint, businesses face greater threats in cyber security. There is a higher chance that a company be the target of cyber attacks compared to consumer homes. PwC notes that there are some limitations and oversight by Boards of US companies, which may influence the use of artificial intelligence to detect cyber attacks. The need for cyber security by companies must be evident before AI can come into play. PwC emphasizes that cybersecurity is an enterprise-wide affair that must be tackled from the top-down, and the compromise of the Internet of Things can be detrimental from global operations, causing not only monetary loss in the cease of production but also physical damage. It can be especially concerning if the source of the attack could be halfway across the globe [22]. Until greater emphasis is placed on cybersecurity, the growth of AI in this sector may be limited.

Future Predictions

In the cyber security realm, artificial intelligence is gaining traction in numerous industry conferences in the United States. Notable companies, such as ForAllSecure and SparkCognition continue to develop cyber security systems powered by AI [23]. Although the uptake of the importance of cyber security may be slower than the emphasis of AI in the cyber security industry, the general trend points to an increasing emphasis on the need for cyber security on both a personal and professional level.

Strange Applications

With such a vast variety of types of artificial intelligence, a few of the stranger applications of AI can be found below.

Robot Prophet - Nautilus

Nautilus:Bin Laden Network [24]

The self-learning supercomputer “Nautilus” was able to analyze millions of documents, and divided them based on two criterias: the nature of the document (positive or negative) and location. The software Nautilus used was able to sift through worldwide news reports. As Nautilus specifically looked at news regarding Egypt, Libya, and Tunisla, Nautilus noticed the trend of negativity of reports that lead to the revolution, “Arab Spring.” As Nautilus anaylzed these documents and news reports, Nautilus was able to accurately find the location of Bin Laden. The same task that took over ten years, two wars, and billions of dollars of the US government, the Nautilus was able to find a 200-km radius that had Bin Laden’s location. Currently, Nautilus is still being tracked events of today to see if it can predict events of the future [25].

Google's Self-learning Computer

Google’s computer scientists created a self-learning computer with a neural network simulation system built with 16,000 computer processors and allowed it to freely access the internet: no restrictions, no rules, and only freedom. In the end, the computer did what many internet users do today - look at pictures of cats. The brain simulation of the neural network simulation was exposed to millions of cat videos and cat products, and began to recognize cats despite no initial information about cats. The system was also able to detect human faces and human bodies at over 76%. The system was first taught to a 15.8% accuracy in recognizing 20,000 objects, which grew to over 70% on its own. The results could prove helpful to language and image recognition software, as well as to how we learn and define language and images ourselves. What this neural system has basically done is understood the concept of a cat. And yet, the system has chosen to do what many web browsers these days do: watch cat videos [26]

Schizophrenic Robot

DISCERN, a supercomputer that functions and operates with the principles of how a human brain functions, functioning as a biological neural network. DISCERN’s is the attempt to recreate the mechanism behind schizophrenia, which follows the concept of “hyperlearning” - to lose connections among individual stories and distinction between reality and illusion. DISCERN is able to use and learn natural language. The scientists taught it a series of simple stories, the same way they would teach it to a human. DISCERN’s role was to learn under the influence of hyperlearning, having a high rate of learning. As a result, the computer was overloaded with many stories. Eventually the computer mentioned many delusional stories, one claiming it was responsible for a terrorist bombing - mixing a third party’s story of terrorists into its own memory. Despite of how creepy this robot sounds, this will benefit researchers to understand schizophrenia to a greater level, improving clinical research [27].


IBM's Watson

In the development of a machine that can effectively understand and process natural language questions while utilizing a network of information and knowledge, IBM developed Watson as an answer to the DeepQA Project. IBM’s Watson is “an application of advanced natural language processing, information retrieval, knowledge representation and reasoning, and machine learning technologies to the field of open domain question answering” [28]. In other words, Watson is a cognitive technology that can understand, reason, learn, and interact like a rational human. It can interpret and analyze large amounts of unstructured data, rationalize personalized recommendations based on personality, learn to become an expert in certain subjects, and interact with Watson through chatbots.

Watson completes the following to reveal insights in data analytics:

  1. "Analyzes unstructured data: Uses natural language processing to understand grammar and context [29].” Approximately 80% of all data is unstructured, coming in the form of tweets, comments, and articles.
  2. "Understands complex questions: evaluates all possible meanings and determines what is being asked,” similar to how a human would rationalize complex ideas [30]. Watson has the speed and capabilities to understand far more complex ideas than the human mind could ever imagine.
  3. "Presents answers and solutions: based on supporting evidence and quality of information found [31].” The recommendations presented are based on the accuracy of the information feeding into the system. They can be personalized for individual use.

From a technical standpoint, Watson is “made up of a cluster of ninety IBM Power 750 servers (plus additional I/O, network, and cluster controller nodes in 10 racks) with a total of 2880 POWER7 processor cores and 16 Terabytes of RAM” [32]. IBM’s DeepQA software is integrated into this network to allow for Watson to process vast amounts of information and tailor it to user needs. Watson is optimized for complex analytics and is currently available via open application program interfaces (APIs) and software-as-a-service (SaaS) products [33].


Notable achievements throughout Watson's life includes:

IBM Watson on Jeopardy [34]
Year Event
2011 Victory on the game show "Jeopardy!" against two champions, Rutter and Jennings [35]
2013 First commercial application would be for utilization management decisions in lung cancer treatment at Memorial Sloan Kettering Cancer Center in conjunction with health insurance company WellPoint
2015 Created 65 new recipes and released in the book, "Cognitive Cooking with Chef Watson," demonstrating ability to pair ingredients together and predict human reaction to taste and texture [36]
2015 Watson Health launch; new global data health cloud that will allow doctors and researchers to share and analyze health data [37]
2016 Ashok Goel, professor at Georgia Tech, used Watson to create a virtual Teaching Assistant to assist students in his class. Jill answered questions where it had a 97% certainty of an accurate answer, with the remainder being answered by human assistants [38]
2016 IBM announced it would be using Watson for weather forecasting. Specifically, the company announced they would use Watson to analyze data from over 200,000 Weather Underground personal weather stations and data from other sources [39]

The numerous achievements of IBM’s Watson shows that AI is capable of absorbing and understanding logical structures with high accuracy. It shows the capabilities of AI and how it can help mankind expand our intelligence or make everyday tasks easier (Cite). IBM Watson illustrates how AI can be used for countless possibilities through using data from millions of sources. It proves how data and keyword-based search engines are not even the slightest achievement compared to how data can be analyzed and personalized through the use of artificial intelligence. IBM’s Watson is, arguably, one of the first to gain public attention of how AI is on its way to becoming more integrated in daily practices and how its possibilities of how it can be used in a wider scope.


IBM Think Academy - How It Works: IBM Watson Health [40]


Watson’s Discovery Advisor system is capable to reading “millions of studies, patents, proprietary documents, and other information” [1]. Compared to a regular researcher, who reads less than 300 scientific papers in a year, Watson is capable of not only comprehensing the data but able to demonstrate connections through data points to make recommendations on medicinal decisions. According to the BAylor college of Medicine in Houston, Watson was able to pinpoint six proteins that alter the protein, p53, after reading over 70,000 studies in a few weeks.

On a similar front, Watson Health, a cloud-based application, is capable of accumulating and analyzing a patient’s medical information from structured and unstructured data. As shown in the video, it understands which medicines a patient may be allergic to and provides recommendations on treatments for a patient. In this sense, Watson can aid doctors in making well-informed decisions, drastically narrowing down unfeasible treatments with supporting evidence [2].

Watson’s reach is not limited to only human patients, but can extend to animals as well. Veterinarians may be posed with the challenge of treating up to 300 breeds of dog and 70 breeds of cat in a single day. By utilizing Watson, Sophie is an application that was created to allow vets access to studies and treatment options for our furry friends [3].


With the use of psycholinguistics, Watson’s Personality Insights program can use people’s writing and word choices to better understand a person’s personality. This can aid businesses in micro advertising and targeted advertising. By automating data preparation and modelling, this can create value for businesses by creating a more personalized method of customer experience. Product recommendations and marketing can be better tailored to audiences or even matching skills with individuals for recruitment purposes [4].


From an everyday individual standpoint, IBM’s Watson was utilized to help military personnel transition back into a civilian lifestyle. An estimated 155,000 military individuals transition back to being regular civilians in a year, presenting a variety of challenges for thousands. Watson and the USAA, a finance firm, allows former military to ask questions and obtain answers to their financial needs [5].

Wayblazer, a start-up built on the foundation of Watson by Terry Jones, links travelers to local areas, food, and offers to get a complete experience in foreign lands. It presents a travel concierge for airlines, hotels, and rentals personalized for the traveler. Using Watson’s cognitive technology, Wayblazer is looking into using natural language search to respond to thousands of queries in less than a second [6].

Business Partnerships

More recently, IBM and General Motors (GM) have developed the OnStar Go system that will deliver personalized offers from partners, for example, to alert drivers who need fuel and to point it to a gas station on route to their destination. It can track the distance traveled, explain complex functional components of the vehicle, and even schedule maintenance appointments to ensure the vehicle is in prime driving condition [7]. Dubbed the “first cognitive mobility platform”, OnStar Go uses machine learning to connect drivers with opportunities and brands in their surroundings. It is limited to the partners chosen by IBM and GM, such as Mastercard, which will allow drivers to pay for purchases through their car’s dashboard [8].

Implications of Societal Integration

Although personalization and targeted advertising can be beneficial by narrowing down the vast amounts of data available in the world of big data, the concern of the “filter bubble” may be brought to light. Eli Pariser coined the term in 2011 as “the way recommendation engines shield people from certain aspects of the real world” [9]. Similar to how social networks personalize information to what an individual tends to agree with, this can cause divisions and narrow-mindedness [10]. For example, if one were to use Watson’s capabilities to find areas to visit for scuba diving and Watson only showed areas according to the search, there may be other activities or landmarks of historical importance that would be filtered out without the individual’s knowledge. Individuals must be willing to be exposed to a wider range of ideas and opinions. In situations such as using Watson for medical treatment plans, the issue of the filter bubble may not be as prominent as the narrowing down of treatment options based on a patient’s medical record may be beneficial for the patient. Using Watson for marketing practices, however, may limit the audience the business is trying to target or the scope and variety of options for the consumer without obvious realization this is happening. In order to avoid the filter bubble, it is important to stay informed and to use a variety of resources, rather than just what is presented neatly in front of you.

Google's DeepMind

“Solve intelligence. Use it to make the world a better place” is the mindset that developed into the vision for Google DeepMind [11] . The company, DeepMind was founded in 2010 in England and was acquired by Google in 2014 for ~$617 million USD [12]. Hoping to use AI to solve complex problems without explicitly teaching it how to, Google DeepMind collaborates with experts from various fields in order to achieve breakthroughs that may have not been possible otherwise. For example, delving into the gaming industry, Google DeepMind was capable in learning and mastering 49 Atari tiles, utilizing pixels as input for self-learning. The AlphaGo program challenged the world champion of Go, a game that is dependant on intuition and complex, logical thinking (there are “more positions than there are atoms in the universe” and came out victorious [13].

By utilizing two different types of machine learning-deep learning and reinforcement learning-it allows for versatility of the Google DeepMind system. Deep learning uses “a brain-inspired architecture in which connections between layers of simulated neurons are strengthened on the basis of experience” [14]. Similar to IBM’s Watson, DeepMind draws on unstructured data to draw conclusions and relevant discoveries. Reinforcement learning was also inspired by how brains function, drawing from the reward centres governed by the transmitter dopamine. Utilizing raw pixels, reinforcement learning utilizes trial and error to determine which actions generate the greatest rewards. Similar to how positive reinforcement is learned by puppies to do tricks for treats, the Google DeepMind system determines which action to take in games, for example, based on positive rewards [15].


Google Deepmind's AlphaGo has beaten Go world champion Lee Se-Dol [16]

Notable achievements from Google's DeepMind includes:

Year Event
2010 Deepmind (company) was founded in 2010 [17]
2014 DeepMind gained traction and was bought by Google for $617 million USD [18]
2015 AlphaGo is victorious in Go against the 3-time European Champion, Fan Hui [19]
2016 AlphaGo continues its victories by winning against Lee Sedol, “the best Go player in the world for the past decade” [20]

Through its achievements, it is evident that Google Deepmind is able to mirror activity similar to that of the human brain, but in a greater capacity of absorbing vast amounts of data and acting accordingly. By winning against world champions in the game of Go, it proves that AI has the capabilities to continuously learn and push towards striving more. It proves that AI is well on its way to learn on its own, and perhaps be translated to apply for more real-world problems. It may find solutions to answers that humans would typically not think of. Rather than a competition between man versus machine, it can be seen as an achievement for mankind, as machines were built by man [21].



Eye diseases or issues are gaining prominence in the world. In diagnosing eye diseases and determining appropriate treatment for these issues, digital scans are used but the complexity of the scans have proven to be a pain point for propower diagnosis. With the help of Google DeepMind, optometrists and eye researchers will be able to gain a stronger understanding of how diseases form and how they can be treated early to prevent effects such as vision loss. Experts from various eye hospitals and charities are optimistic in utilizing DeepMind’s AI in fighting eye diseases and researching towards early prevention [22].

Energy Savings

Utilizing the AI behind Google DeepMind, Google was able to reduce emissions and increase energy efficiency in Google’s Data Centre, allowing cost savings of 40% of its current levels. Utilizing neural networks of Google DeepMind, historical data collected from existing sensors in the data centre allowed Google to determine the average future PUE (Power Usage Effectiveness), “which is defined as the ratio of the total building energy usage to the IT energy usage” [23]. With two other neural networks, DeepMind is capable of predicting future heat and pressure trends and recommend next steps within operating constraints. Going forward, this type of analytics can be passed on to “improving power plant conversion efficiency, reducing semiconductor manufacturing energy and water usage, or helping manufacturing facilities increase throughput” [24].

Implications of Societal Integration

As with IBM’s Watson, the filtering and narrowing down of information can induce the effects of the filter bubble. Currently, Google’s DeepMind is in its early stages of practical uses compared to its current achievements in gameplay. However, as mentioned earlier, the success of Google DeepMind proves AI’s capabilities and potential in a variety of real-world applications in the future. With this also comes the “fear” over AI and its ethical influences, discussed in the #ETHICAL CONSIDERATIONS section.

Salesforce's Einstein

Salesforce, as one of the most popular Customer Relationship Management (CRM) tools based in the cloud, is one of the first to offer AI services in the CRM realm. While a relatively new offering, Salesforce attempts to remove the technical limitations of installing AI in existing infrastructure. Powered by a series of AI types, including “advance machine learning, deep learning...and predictive analytics,” Einstein is integrated into the CRM platform. It claims to be “automatically customized for every single customer, and it will learn, self-tune, and get smarter with every interaction and additional piece of data” [25]. In other words, it provides further personalization and recommendations that is commonplace with AI platforms today. Salesforce claims to provide intelligence geared towards business models, which will help with automating tasks for sales people and determining which leads are most likely to result in a closed deal so salespeople can direct their efforts accordingly.

Similar to IBM’s Watson and Google DeepMind, Salesforce Einstein attempts to discover and validate insights, predicting outcomes based on historical data and recommend what steps should be taken next [26].


As a relatively new “product”, Salesforce Einstein has mixed reviews on its usability and usefulness for the business. Salesforce Einstein is the result of their research into AI for two years. Rather than a new standalone product, Salesforce Einstein is integrated into their CRM model. It can be viewed as a “set of intelligence functionality that underlies the entire Salesforce platform...the idea is to provide a base on top of which the company can continue to add new capabilities into the future” [27]. By building the solid foundation of AI, Salesforce can combine data and computing together for future growth. In this sense, Salesforce is looking towards the future and the capabilities of AI, rather than shy away from the idea.


Salesforce Einstein Applications [28]

Salesforce Einstein is currently offered on these platforms, each tailored to a different audience for varying purposes:

  • Sales Cloud Einstein
  • Service Cloud Einstein
  • Marketing and Analytics Cloud Einstein
  • Community Cloud Einstein
  • IoT Cloud Einstein
  • App Cloud Einstein
  • Salesforce provides information on how each of these cloud Einstein’s can utilize AI to develop greater results[29]. For example, the Service Cloud Einstein offers insight on how to respond to customers and empower employees to enhance customer service across various realms. A product with a known defect can direct a service agent to the appropriate means of escalation and resolution based on previous actions of similar cases[30].

    Implications of Societal Integration

    Salesforce Einstein was introduced in September 2016, making it a relatively new feature in the AI space. Its capabilities are quite limited right now, but the fact that Salesforce has developed a team dedicated to AI and are forward-looking illustrates that larger companies are willing to take steps in unknown and controversial directions [31]. Salesforce Einstein is an open AI interface, allowing for collaboratively efforts for developers for experimenting. Salesforce is arguably one of the strongest leaders in the CRM space: furthering their business platform based on AI proves the sentiment that the future will move towards an AI space, whether the general population agrees with it or not.


    Job Replacement

    Job Replacement Varies by Country [32]

    A common conception of AI is that it will become more useful than humans in the long run and eventually replace them. The concept of automation and how it replaces jobs occupied by humans has been a topic of controversy for years, so the undefined capabilities of AI that mirrors the brain’s neural networks can be nerve-wracking for many. While the more pessimistic claim that “77% of jobs in China will be vulnerable to robots or AI replacement” and 40% of jobs in the banking industry could be replaced, there needs to be the existing infrastructure and government regulation to support this type of advancement [33].

    Furthermore, although simpler manual jobs may be automated, new jobs can be created with new technologies as well. It can develop different employment models or create jobs that may not have previously existed, depending on the time and demand. For example, social media coordinators or YouTube as a full time job never existed previously, but developed with the times as social networking and smartphones become more commonplace. AI can help stimulate more jobs in the marketplace and implore new skills that may not have been necessary before. According to a study by the International Federation of Robotics, employment may increase in the fields of electronics, automobiles, renewable energy, and even food and drink [34]. While the introduction of AI and automation may replace certain jobs, it can create greater efficiencies and business models to promote job growth in different methods, inducing change that many may be afraid of. These learning machines can be viewed as helping humanity, rather than replacing it as mentioned with Google DeepMind’s achievements. AI has the potential to aid humans in adapting to rapidly changing issues, environments, and systems, while man can provide creativity and genuity that technology may not be able to mimic.

    Human Dignity

    Positions that require respect and care, such as judges or doctors, are under debate on whether or not they should be automated through some form of artificial intelligence. These jobs that require human judgement and care, some argue, will not be able to be automated through AI. Machines should not make decisions on humans, especially in situations where humans may be put in danger or captivity (encroaches human dignity). Positions, such as doctors, that require empathy may lead to frustration and isolated feelings for patients that are going through difficult experiences in their lives [35]. More likely than not, artificial intelligence will rely on historical data and precedence in determining sentences or treatments for humans, which may not always result in a “fair” result in the grey areas of justice and ethics. The overall sentiment regarding AI in these fields is that humans require the social and empathetic components of such vulnerable environments, making AI a topic of hot debate. In these cases, it is safer to say that AI should be used to aid individuals in making more informed choices, rather than making the choice itself.

    Artificial Bias

    Although a benefit of artificial intelligence is its ability to make unbiased decisions, recent studies have shown otherwise as it is often overlooked that AI is created by humans who, themselves, have biases. AI has been under fire for being sexist and racist particularly in facial recognition during two famous incidents where Google's photo application automatically classified black people as gorillas [36] and when Nikon's camera software mistook pictures of Asian people blinking [37]. A recent study has also found that women were less likely to appear on Google ads for highly paid jobs than men. This could be the bias of advertisers or simply the unintended consequences of the algorithms used by the search engine [38]. It may be just a data problem but fundamentally, the algorithms which these programs run on are fed images by engineers who play a large part in how the system works and identifies different objects and environments.

    Face detection in photos may seem like a trivial problem compared to the most criticized AI software, COMPAS, which was created by Northpointe Inc. to determine risk scores for offenders. It was designed to help predict which defendants were more likely to commit new crimes and therefore aid the criminal justice system in assigning the terms of incarceration. A system that could correctly identify and predict these offenders was the goal but the opposite happened and two significant discoveries were unveiled following the analysis of ProPublica:

    • "Black defendants were often predicted to be at a higher risk of recidivism than they actually were. Our analysis found that black defendants who did not recidivate over a two-year period were nearly twice as likely to be misclassified as higher risk compared to their white counterparts (45 percent vs. 23 percent).
    • White defendants were often predicted to be less risky than they were. Our analysis found that white defendants who re-offended within the next two years were mistakenly labeled low risk almost twice as often as black re-offenders (48 percent vs. 28 percent)" [39]

    COMPAS' algorithms are hidden and these skewed results have raised many concerns over the reliability of related data-driven-risk-assessment tools for predictive policing. Some have pointed out how the software promotes a vicious cycle as the increase of police presence in certain areas almost guarantees greater arrests in those areas. Particularly, in the United States this poses a problem as greater surveillance in poorer and nonwhite communities will always we under more scrutiny than wealthier and whiter neighborhoods. In the end, predictive systems and programs are a reflection of the data they are inputted and trained with.

    Privacy and Security

    In order for artificial intelligence to improve and be more relatable to consumers and businesses alike, data collection is inevitable. For example, with IPAs such as Apple’s Siri and Google Home, millions of voice clips are recorded and stored in servers indefinitely. From recording your address to potentially manipulating someone’s voice to obtain sensitive information, this may change the way consumers interact with businesses on a number of fronts. For example, most credit card companies either use your voice and sensitive information in order to verify your identity. If a group with malicious intent obtained your voice clips and was able to recreate your voice, they could easily bypass this identity verification security step.

    The concept of encryption and security has been a topic of discussion for decades; in terms of AI, although there is an option to encrypt certain messages, the AI component is not compatible with encryption[1]. In a White House report released a few months ago, recommendations are made on ensuring innovative methods are used to secure artificial intelligence in terms of cybersecurity[2]. Privacy and security of information is an ongoing issue in the whole realm of technology and electronic advances. The right questions have to be poised in order to ensure security of information is pushed for, especially towards the larger companies that have the means to secure their systems that millions use each day. Cybersecurity will continue to gain traction in the years to come.

    Robot Rights

    The question is this: “Do robots that exist human-like behaviours but are essentially technology entitled to having rights, similar to humans?” A grey area that comes to light is how we define the humane treatment of artificial intelligence as they become more “human-like”, with their own personalities and capabilities of social interaction. The EU previously discussed the introduction of new laws for robots and providing them civil rights, similar to human beings. The “electronic person” would be held responsible for its actions and also have various rights[3]. The most prevalent conclusion would be that robots do not have the same level of rights as humans, as they do not necessarily have emotions or feel pain the same way. This also brings into question, however, how robots would be “held responsible” for their actions. Would it even be a punishment if say, a robot were “switched off” due to an irresponsible action? If anything, it may be a step backwards as the artificial intelligence wouldn’t be able to learn from its actions. These difficult decisions will soon, if not already, come into the radar of governments.

    Moral ethics of human-turned robots (for example, cyborgs) complicates this issue further. Ray Kurzweil, a world-renowned scientist in his experiments and expertise in AI, believes that we may one day be able to upload various components of our brains into the cloud. AI would be linked to humans to form ultimate computing machines. It takes the computing ability of the human brain away from its biological constraints[4]. This specific concept could make humans cyborgs, which brings robot rights and human ethics deeper into the grey areas. Although these scenarios may seem unrealistic and impossible, our current growth and advances into AI may prove otherwise. There is no clear (and there may never be) to whether or not robots or cyborgs should have rights, but it seems that popular opinion is that robots do not deserve rights while cyborgs may be entitled to some (as they are essentially still human).

    Outsmarting Humans and Malicious Intent [5]

    Tay.AI was AI created by Microsoft that existed as a chatbot on Twitter. The bot would speak like millennials, learning from people it interacted with through social media. However, it took less than 24 hours for Tay to turn 180 degrees from “Humans are cool” to “Hitler was right” through its learning[6]. Some argue that this shows Microsoft’s success in creating a bot indistinguishable from humans and how relentless social media platforms can be. Continuing on this train of thought, Google’s AI was able to create cryptography between two neural networks that was indistinguishable by the third. This may mean that even humans or other AI systems may not be able to decrypt in the future, “outsmarting” humans with a mind of their own[7]. If AI were to “turn” against humans and showed malicious intent, what steps could man take to overcome this? AI can be used as weapons, either between humans or between machines and humans.

    Government Regulation

    Governments by far have the largest influence on a number of factors, bringing into difficult decisions for lawmakers. As defined earlier, there are many grey areas that need more clear definition through government regulation and decision-making. Should self-driving cars save drivers or minimize impact on outside factors? What if AI lead to a wrong diagnosis and lead to death? All of these questions are posed without clear answers, similar to how there wouldn’t be clear answers if it were humans in that situation. However, since AI is not human, we do not perceive it in the same light.


    One viewpoint is that AI can eventually and theoretically take over the world; ending humanity in the future if not careful. Elon Musk feels the inclination that AI requires a regulatory oversight on national and international levels to avoid foolishness. Stephen Hawking explains that AI has proven itself useful, and thorough development of artificial intelligence could end the human race. Theoretically, artificial intelligence could eventually become independent and “redesign itself at an ever-increasing rate" [8]. Bill Gates also stated on Reddit that he agreed with Musk and that artificial intelligence should be a concern [9]. In general, the current AI community as a whole still requires much development and research before building anything that could be a strong concern of robot takeover to the general public. The reality of AI research and development is complex and the creation of human-level AI is a difficult if not impossible goal.

    There is a certain distinction between soft AI and hard AI, where hard AI acts like the human mind. Soft AI is used in the everyday lifestyle, assisting with tasks that are generally statistically oriented and computational methods are addressed using analysis and algorithms. Such examples are airline reservations, refilling prescriptions, and instructions from the GPS[10].

    Hard AI is made with intelligence in cognitive tasks mainly revolving around reasoning, planning, learning, vision, and natural language. The intelligence levels are ones that can match or exceed human intelligence, and some believe it is a consequence of Moore’s Law – where the number of transistors incorporated in a chip doubles about every two years[11]. Once AI exceeds human intelligence, perhaps machines may display a consciousness or self-awareness, experiencing sensations and feelings. Bill Gates believes that in “a few decades after that though the intelligence is strong enough to be a concern” [12]. However, Gates also believes that the growth of technology helps people get more questions answered, allowing the masses to understand more topics and contribute together to solving complex problems, leading to extended debate over artificial intelligence.

    The Growing Future of AI

    MITCSAIL - Action-Prediction Algorithms [13]

    While the possibilities of AI are endless, there are several trends that are gaining more traction. In particular, a few important concepts include:

    • With the rise of intelligent personal assistants and a focus on convenience and ease of use, there are trends towards the use of zero interface. People are slowly shifting away from using interfaces on screens "in favour of speaking directly to faceless machines" [1]. On the business forefront, brands must look into alternative methods of advertisement and communicating with both current and future customers[2]. As with the transition of newspaper ads to online and email marketing, it may not be long before there needs to be a transition from online marketing to something not yet defined.
    • Another growing area of AI lies with making everything "smarter", whether it be as simple as a kitchen appliance to the traffic control of a whole city. A study presented recently identifies the opportunities of smart infrastructure: "a distributed lighting system to facilitate the implementation of anew infrastructure in a new city" and build on existing infrastructure to reduce energy costs and optimize energy usage. This may also benefit cities in environmental sustainability and efficient resource allocation [3].
    • Perhaps one of the greatest goals of artificial intelligence is the ability to predict the future based on historical trends. A group of researchers from MIT was able to create a deep learning algorithm to predict subsequent events based on a single video clip, as seen in the YouTube video to the right. The AI system used patterns and social signals to predict what would happen next, which is an indicator that maybe in the far future, AI will be able to predict actions which can be translated to avoiding nuclear warfare or climate change [4].
    • One increasingly common idea in the technology are Smart Cities, where balancing the environment and saving natural resources is the ideal goal. Setting lighting schedules for all public controlled lighting installations for energy savings is achieved by regulating light intensity. Defining their own light patterns, refers to the hours during which the lights are on, and what level of brightness, and assigning different light patterns to each day. User preference and many other factors come into play, such as weather, traffic, and astronomical clock. And by combining AI techniques like artificial neural networks, EM algorithm, methods based on ANOVA and a Service Oriented Approach, can create an innovative system in intelligent prediction of consumption and cost. This allows light patterns to be predicted easier even with sudden pattern changes [5]. People can leave their houses without worrying about turning off lights or appliances, and can expect their home to start warming up as they head home on cold days.
    • With the constant growth of technology, so does its capabilities in assisting people with learning and growing. With the use of AI, education and people’s ability to learn can be enhanced even further. Using AI can improve or change the current education system, and can become more accessible to areas where education is unaffordable or scarce. AI can automate basic activities in education, becoming more efficient. AI can also adapt educational software to the students’ needs, improve course material and offer as support. Improving the learning environments can achieve better results, but AI is not limited to such capabilities. There could be AI-driven programs that can give students and educators helpful feedback, change the role of teachers, becoming the teachers, using data supported by AI to maximize the potential for schools to assist and teach students, and maybe even create stronger environments of learning [6]. AI tutors and teaching assistants already exist, and has shown learning efficiency. The potential of assisting learning and growth can be endless, and with proper use of AI can improve the level of intelligence of humans worldwide.

    Our Thoughts

    A few years ago, it seemed like aritificial intelligence was something out of a science-fiction movie; something that would come to life hundreds of years from now. However, AI is here right now and many considerations, both ethical and regulatory, need to be taken into account in order to successfully integrate it into our society.

    Our predictions are that AI will coexist and assist humans in their everyday lifestyle in the future. The idea of creating powerful AI without it turning against us is achievable. Rather than having artificial intelligence replacing jobs, acting maliciously, and overpowering humans, AI should be carefully regulated and used to aid mankind in its endeavors. As a result, a new lifestyle where jobs and services are assisted with AI can be created, while new levels of AI can assist in our own intelligence and capabilities. We can turn machines into workers: they can be thought of as labour, but this greatly undercuts human value and potential... something that automated machines may not necessarily have. Before definitely making the statement that AI may replace humans in the future, we need to properly define the way society values people and their potential, not just by the amount of productivity or work that they can complete efficiently. Humans need to value each other first, and that machines should be used to benefit people. Continuing on this mindset, this can lead to AI and human teams to answering and solving the world’s most pressing and complex problems, such as climate change, conversation and distribution of resources, controlling diseases, solving geopolitical problems and even fostering world peace. The question is not, "who will win between man or machine," but rather it should be: "how can man and machine work together to tackle the world's largest problems in new and innovative ways... how can we achieve the impossible?"


    1. Bacon, J. (2016). Why marketers should take notice of Amazon and Google's connected home launches. Marketing Week (Online Edition), 5.
    2. Bacon, J. (2016). Why marketers should take notice of Amazon and Google's connected home launches. Marketing Week (Online Edition), 5.
    3. De Paz, J. F., Bajo, J., Rodríguez, S., Villarrubia, G., & Corchado, J. M. (2016). Intelligent system for lighting control in smart cities. Information Sciences, 372241-255. doi:10.1016/j.ins.2016.08.045
    4. Cuthbertson, Anthony. (2016, June 22). Artificial intelligence algorithm predicts the future Retrieved from Newsweek:
    5. Juan F. De Paz, J. B. (2016, August 17). Intelligent system for lighting control in smart cities. Retrieved from ScienceDirect: 2016
    6. Writers, S. (2012, October 30). 10 Ways Artificial Intelligence Can Reinvent Education. Retrieved from Online Universities:
Personal tools