Artificial Intelligence

From New Media Business Blog

Revision as of 05:58, 6 January 2016 by Rchase (Talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

Contents

Artificial Intelligence

What is AI


The theory and development of computer systems able to perform tasks that normally require human intelligence. Artificial Intelligence (AI) can take many forms and has the ability to placed in anything that has a computer processor. It has a variety of potential uses from completing household chores that we humans do not want to do to more complicated procedures such as helping diagnose patients.



Purpose

AI originated from scientists wanting to make complex decision making and management based computer systems. Now, AI’s main purpose is to be integrated into society as a tool to make lives easier, such as the invention of a computer did.

Forms

The two mains forms of AI are agents and multi agents. Agents repetitively complete a single task [1], whereas multi-agents are able to interact, learn via feedback, and have semi human characteristics [2]. The following are examples of agent and multi-agent systems created by RethinkRobotics lab [3]:

Agent

How Roomba Works

The Roomba Vacuum series is a vacuum that is small, portable, wireless, and efficient. It does not require a human operator to do manual labour. Once turned on, the vacuum will measure the room’s size, shape, contents and flooring. The Roomba is able to manoeuvre around objects, including furniture and pets, and clean the floor before moving onto the next room. The Roomba will only adjust the pattern in which it vacuums the room, not how it vacuums.



Multi-Agent

How Baxter Works

Baxter is one of the first cost-effective multi-purpose robots and computer systems produced. The video will illustrate how it works. Baxter was programmed to accomplish a variety of things, such as folding shirts, moving stuff, and organizing. Baxter’s best feature is his ability to learn and adjust his technique each time he completes a task.




As technology becomes more affordable and developed, AI will be further integrated into our society. For the moment multi-agent and single-agent technologies can be easily used in the service industry and manufacturing to complete jobs that humans prefer not to do. Thus, AI will allows humans to take on larger and more ambitious projects.

How it Works

Machine learning is used in AI to improve and learn the same way humans do. Machine learning is based on algorithms, epistemology, computer science, biology, and probability. It is the application of statistics and mathematics to teach computers to learn, think, and predict [1]. Although AI uses research and knowledge from a variety of academic fields, AI would not have been able to achieve the same advancements without machine learning.

History

AI is a fairly recent concept that emerged during the research of nuclear bombs in World War II and Allen Turing’s complementary research. The following table is a timeline of major AI research breakthroughs and links for further information about them:

Major AI Events
1950: Alan Turing’s Turing Test [2]
1956: Dartmouth Coined term AI [3]
1974-Early 1980s: Winter of AI [4]
1987-1993: Second Winter of AI [5]
1997: IBM’s Deep Blue Computer [6]
2005: Darpa Grand Challenge [7]
2011: IBM’s Watson [8]
2012: Google Brain [9]


Allen Turing created the ‘Turing Test’ to understand machine learning. He conducted an experiment and analysis to decide whether the behaviour exhibited by a machine is equivalent or indistinguishable to that of a human. The test consisted of a game called, ‘The Imitation Game’, where the interrogator tries to guess the gender of the other two players. For a machine to pass the game, the interrogator had to guess the machine was human more than 30% of the time during a series of five minutes written conversations. Machines did not pass this test until June 9, 2014 [10]. Vladimir Veselov, Leobner Prize Winner and expert in conversational systems created a computer program called Eugene Goostman, a 13-year old Ukrainian boy. He took the Turing Test with several world renowned experts in AI and passed by convincing them that Eugene was human 33% of the time.

In 1956, the term “Artificial Intelligence” was coined at a conference in Dartmouth. This was a significant moment for AI because it defined its field and gave it opportunity; however, this period was followed by a stall in research referred to as the first AI winter.

In the early 1980’s, technology caught up to the goals of AI and researchers were seeing results in their advanced systems that attempted to replicate human intelligence. Majority of the funding to advance AI was provided by the government, but this stopped after a couple of years and created the second AI winter.

Eventually, the second AI winter ended when IBM’s Deep Blue Computer beat the undefeated chess champion of the world, Garry Kasparov. This breakthrough was followed by the Defence Advanced Research Projects Agency [11], which created an autonomous self-driving car named Stanley. Stanley used machine learning and probabilistic reasoning to navigate through rough terrain without a driver. At this point in time, the technology finally matched the research goals of AI.

The closest system that resembles the AI of the future is IBM’s Watson. In 2011, IBM created Watson to play against world champions in Jeopardy. Watson’s software included custom IBM software and an unstructured information management architecture system to start the process of machine learning so Wason could sort through the information once questions were asked while playing Jeopardy. During the show, Watson had no internet connection and was able to answer all of the questions correctly. Today, IBM has found more practical business applications for Watson, such as medicine, consulting, and analytics [12].

The latest major breakthrough is a branch of AI that researchers have been developing since the early 2000’s. Deep learning, an algorithmic based machine learning technique, has been applied to AI [13]. This is another layer of necessary depth that will allow machines to further learn, think for themselves, and make decisions with the same level of consciousness as humans. Currently, deep learning is still in its early stages of development.

To conclude, AI is a relatively new concept that has been developed by a multitude of people from a variety of professional industries. The history of AI is based on a collection of knowledge and breakthroughs that teams of researchers have put together to create the AI we use today.

Modern Applications

Currently, AI is being used today but because it is mostly agent based technology it benefits from not being in human like form. These forms of AI that we will interact the most with are elaborate service systems. These systems are used to confirm our knowledge and to ask questions, to drive us places, and to do the basic activities we feel we do not need to do; which makes it easier for them to be accepted when they take our jobs.

Transportation Industry

Public Transportation
Subways and skytrains are examples of autonomous self-driving public transportation. These forms of transportation can maximize efficiency and lead to increased revenues by allowing the optimal number of trains to run on the track without manual labour to restrict its operations [14]. The weaknesses of this technology include its high-cost of investment for construction and its potential to increase project risks. The technology is also at risk of being outdated and un-adaptable to new technological advances. Although the trains do not require a driver, the system still requires human supervision and possible intervention to make ethical decisions about system failures and maintenance costs. As AI continues to advance, there is a possibility to make the system fully automatic and to program it to make decisions in ethical situations.

Beyond public transportation, AI is becoming a growing force in autonomous vehicles. Using AI-powered supercomputers to create the main engines, self-driving cars (a.k.a autonomous vehicles) are no longer an idea of the future. The National Highway Traffic Safety Administration (NHTSA) and companies like Google, Mercedes, BMW, and Tesla have already released or are in the midst of releasing features that give the car some ability to drive itself.

This drastic change in how we drive cars will be revolutionary. Futurists estimate that there will be dire consequences for businesses who do not adjust fast enough [15]. These technology laggards could lose hundreds of billions of dollars in revenue. This report will describe how self-driving cars will affect businesses and our society.

Car Manufacturers
Major automakers will see an initial surge in revenue due to new and used car sales, but sales could decrease significantly as technology takes over and the sharing economy becomes more popular. Cars will continue to need steel, glass, an interior, a drivetrain and some form of human interface, but everything else has the potential to change. Futurists argue that cars will no longer be made of heavier gauge steel and eight airbags if accidents will become rare. Automakers who are well aware of the potential changes are focusing on services as much as they are on what and how they manufacture.

Infrastructure Transformation
With fewer cars, parking lots and spaces that cover many of our city’s open space can be repurposed. This transformation could indicate temporary downward pressure on real estate values as supply increases and it could indicate an increase in greener urban areas. Fewer cars could also lead to the potential of less government spending and maintenance on highways.

Oil Demand
Self-driving cars should practice very efficient eco-driving practices. As well, there could be a reduction in oil demand if people share autonomous cars.

Safety
Automakers are placing more emphasis on systems that will monitor a car’s surroundings, warn the driver of danger, and even take control of the car in some situations. The benefits in terms of safety and convenience could be significant as research shows that 90 percent of crashes are caused by human error [16]. Although it is too early to quantify exactly how self-driving vehicles will affect insurance rates, futurists question who should be held responsible for any kind of accident: the vehicle manufacturer, the software company who designed the autonomous capability, or the “driver”.

Other Transportation Options
Self-driving cars could have a substantial impact on the taxi and limousine industries and public transit companies, but they could also create new ones. Self-driving cars could be used to share specific trips as a pay-as-you-go small-scale version of public transportation. These cars of the future offer as much convenience as rail service with the added convenience that the service can drive you to precise destinations instead of stations. As well, self-driving cars have the potential to save you an airfare by driving you to your destination while you sleep in the car overnight.

While these cars offer many benefits, there will be many regulatory and legislative obstacles as well as security and privacy issues to overcome. However these situations may conclude, the revolutionary technology is here. The full adoption will take decades, but the convenience, cost, safety and other factors will make them ever-present and indispensable.

Manufacturing Industry

AI robots in manufacturing, specifically factory or assembly line work, would drastically reduce costs, increase efficiencies and give opportunities for economies of scale [17]. While the implementation of AI in factories may produce high initial investment costs, it provides manufacturing companies with the opportunity to opt out of using unethical labour sources overseas. These machines are assist the humans workforce or completely automate the system, potentially reducing errors and the need for supervision. AI will also increase savings for businesses by making the workforce more efficient and it will provide returns on investments more quickly. Currently, factory workers make up a large segment of the overseas labour force. By removing these employment opportunities, unemployment rates will increase. Societies revolutionized by AI manufacturing would have to restructure their labour force and potentially retrain their population. On another note, full AI automation in manufacturing has great potential but the costs to society at the moment are greater than the benefits.

Health Industry

Bright.md
AI is starting to be utilized in medicine in the form of software tools to make doctors more accurate, efficient, and effective. For example, Bright.md is a software that is available online and via a mobile app that helps improve physician efficiency by automating their work [18]. By helping doctors treat a patient, write prescriptions, and schedule follow ups, Bright.md enables doctors to increase their capacity for volume and save time and costs simultaneously. Bright.md also decreases the waiting time and a patient’s exposure to other illnesses in the waiting room [19].

Before speaking to an actual doctor, Bright.md works by prompting a patient to complete a smart exam to gather basic data. The artificially intelligent system adapts a series of medical questions to the given answers to ensure that the exam the patient fills out is best tailored to their symptoms. Once this basic data is gathered and stored, the system sends the patient’s doctor a preliminary diagnosis and a treatment plan [20]. Bright.md is designed as a medical tool to assist doctors. Its goal is not to replace doctors and remove human interaction entirely, as human interaction is the cornerstone of medical treatment. The Bright.md system also automatically makes a doctor’s job more efficient by generating documents such as chart notes and insurance coding. By automating the repetitive parts of the doctor’s job, AI systems like Bright.md can enable the doctor to focus on treating the patient.

Potential issues with Bright.md include the possibility that patients do not fully understand the smart exam and fill in incorrect information regarding their illness and symptoms as a result. Though the patients are encouraged to see the doctor after completing the smart exam, the purpose is that the preliminary questions have already been covered so the doctor would not have to repeat them; however, there is a possibility of misdiagnosis if the information provided is incorrect.

Electronic Medical Records
In modern times, a doctor is often found spending more time writing notes than being face to face with the patient. Doctors are constantly multitasking while attending to their patients due to the increasing demands of their profession. Multitasking may increase a doctor’s chances of missing key information or making errors in data entry, which may lead to misdiagnosis. The use of AI systems like Electronic Medical Records (EMRs), Apple’s Siri, and IBM’s Watson could solve these issues by increasing the amount of face to face time doctors spend with patients and reducing the potential of human error.

EMRs are digital versions of a patient’s medical history that allow doctors to view accurate and up-to-date information on their patient’s medical history and health details. These records are used extensively throughout present day medical practices [21]. Vendors, such as TELUS, have recently taken a keen interest in the field of EMRs [22].

On another note, patients could engage in conversation with Siri to assist with the medical process. For example, the conversation would sound like this [23]:

Dr. Stone: "Siri, I would like to admit Ms. Jones to the hospital for her knee replacement."
Siri: "Sure, Dr. Stone. Shall I use your pre-op order set?"
Dr. Stone: "Yes."
Siri: "Tell me the medications she is on."
Dr. Stone: "Okay, let's be sure to let her cardiologist know to adjust her blood thinners a few days before surgery. And by the way, the medication she is on has been recalled and this alternative is recommended."


This conversation demonstrates just a fraction of the possibilities that AI could have on the use of medical records. Other benefits include the following:

  • AI could save medical staff time by providing relevant information quickly and easily
  • Nurses could dictate information from their rounds then AI could be used to notify the doctor if something is concerning
  • Doctors could spend less time entering data into the EMR and could simply dictate it to the AI to record and document

The use of AI in medicine could prompt the doctor to consider a less expensive alternative to a drug, a dressing, or a therapy. Thanks to a simple conversation with an AI system, doctors could have the knowledge base of supercomputers to make their differential diagnoses.

IBM’s Watson
Watson is an IBM supercomputer that combines artificial intelligence and sophisticated analytical software. To replicate a high-functioning human’s ability to answer questions, Watson accesses 90 servers with a combined data store of over 200 million pages of information, which it processes against six million logic rules. The device and its data are self-contained in a space that could accommodate 10 refrigerators. Watson is not programmed with all the information it might need, but instead it has been given the cognitive tools necessary to acquire the knowledge itself, teasing out answers to complicated questions from vast amounts of electronic information. It does this using natural language (everyday conversational English), not computer-language queries. Watson was first introduced when he won a game of Jeopardy against two Jeopardy champions in 2011 [24].

Watson was initially used in the medical field at the Memorial Sloan-Kettering Cancer Center [25]. The doctors asked Watson to learn all of the information it could about oncology and the case histories at Sloan-Kettering. With Sloan-Kettering treating more than 30,000 patients annually, Watson takes information about specific patients and matches it to the huge knowledge base incorporating published literature and the treatment history of similar patients. Watson’s ability to mine massive quantities of data means that it can also keep up with the latest medical breakthroughs reported in scientific journals and medical meetings. Additionally, using cognitive computing, Watson continually “learns,” thereby improving its accuracy and confidence in the treatment options it suggests. Watson is able to prompt doctors if any information is missing and its goal is to display several choices for the physician with various degrees of confidence and to provide supporting evidence from guidelines, and published research.

Essentially, Watson has the capability to learn about specific patients and create a personalized ranked list of treatment options. This information offers a second opinion for the consulting doctor as Watson is able to sift through an immense amount of data and information to make the suggestions.

The implications of using a supercomputer, like Watson, in the field of medicine are large. For doctors, the potential to access technology that can scan the internet and look at the latest medical publications, journals, and breakthroughs before providing possible diagnoses is groundbreaking. Watson is able to do what any one human being could not even fathom with the amount of development and new ideas that come about daily. For doctors, using AI as a tool to support their diagnosis could lead to lower rates of misdiagnosis and more effective and modern treatment plans. Though the current state and use of AI is nowhere near replacing doctors as a whole, the support it can provide can make the jobs of doctors much easier and more efficient. IBM has created a business unit for Watson specifically to see how the technology can grow and be used in a variety of fields.

The risk of using technology like Watson in medicine is overreliance. Doctors currently undergo years of education and practical hands on work before becoming full fledged doctors. The field of medicine is based on sharing ideas, studies, and passing down knowledge through conferences, studies, and networking. With technology like Watson, or even the general use of AI in medicine, the concern becomes whether or not doctors and medical staff will become so dependant on technology that their own learning and expertise will be irrelevant. This shift could allow the possibility for medical staff to focus more on research and expanding the body of knowledge, while the technology can handle the operational tasks. Going further, the question becomes “Will AI eventually replace the need for human beings to be active in the medical field as experts or will the role of humans be to carry out tasks instructed by technology?”

Crime and Surveillance Industry

AI plays an important part in reducing crime. The purpose of the surveillance camera is to capture a live video recording of an area and store the recording on a disk for proactive purposes. This technology helps police track suspects by matching individuals to the recorded footage; however, matchmaking is often done manually and there is a major delay between the crime occurring and the footage monitoring.

The advantages of this scenario is that AI can catch minor details that humans might miss [26]. If AI systems are programmed to learn, they could act a certain way without succumbing to human needs. Once this technology evolves further, it would be more difficult for criminals to commit crimes since the AI surveillance can correctly identify criminal activities. Furthermore, the evolution means criminals can no longer count on humans negligence to the surveillance cameras to get away with their deeds.

Andy Baker from the National Crime Agency made a software that can trick child predators, identify them and directly incriminate them in a court of law [27]. This method of surveillance is superior since police have difficulty getting proof and catching those people without these systems.

Another example is an airport security system’s ability to single out people considered high risk. As technology improves, some people will try to find a way around the new security methods. Privacy Visor Glasses from Japan can jam surveillance cameras by covering the user’s face a white screen [28]. While this technology creates a white screen now, it is not difficult to imagine technology improving to the point where the user could project a different face instead.

The aforementioned examples are basic types of AI surveillance. Using heuristics with drones, these advances to AI could create the ideal policemen. Heuristics is a part of machine learning, whereby a machine uses past and present data to form a conclusion and it learns by putting the conclusion into the database of facts to form more conclusions. Hitachi has claimed it has successfully created a machine (Predictive Crime Analytics) that can predict crimes before they actually happen using patterns in criminal activity to predict where the next crime will occur [29]. This machine uses data from past crimes, social media, weather, etc. to help it predict the location and time of the crime. This technology is different from other machine learning devices, as this machine operates independently of human intervention. By removing human intervention, the machine reduces the chance of human bias, such as racism or sexism. If this technology proves to be accurate all the time, police could send a squad to the place where the computer thinks a crime might occur to prevent the crime. The potential of this idea could decrease crime rates. On the contrary, people might try to change their criminal activities to alter their crime patterns and confuse the machine.

At this time, police cannot arrest anyone for crimes they have not committed because of social and ethical implications and laws; however, this technology may become beneficial to companies and society in the future by preventing the occurrence of crimes and reducing the cost of damages caused by criminal activity.

Challenges

Threats to Society

Professor Stuart Russell, who leads research on AI, believes that AI might be weaponized in the future [30]. He believes that placing weapons systems under the control of a machine, rather than a human, is extremely dangerous. Even Stephen Hawking and Elon Musk have written an open letter to the United Nations asking them to be conscious of what AI could mean to society if used inappropriately [31]. Since AI systems do what they are programmed to do without question, ethical reasoning, or remorse, there is a potential to use them as weapons.

Over the decades, science fiction has increased the fear of AI systems in humans. For example, the American science fiction franchise, Terminator, showed the severe consequences of AI becoming weapons. In these films, the computer programmer programmed the computer to destroy military-based threats aimed at the United States of America (USA). In a turn of events, the machine determined that humans are threats to the USA and, as a result, the machine continued to obey its programming code by attempting to eliminate the human race.

Stephen Hawking also said that we only ever need to make one fully intelligent machine, as that machine would be able to redesign and improve itself at a rate faster than we can improve ourselves [32]. As such, humans would not be able to compete, and they would be superseded by the computer. This would be the end of human race. Consequently, human values and objectives must remain central to the development of AI to prevent a catastrophe. Once a machine has been created for the wrong purposes, everyone would pay the price, not just the creator.

Limitations

Expendability of Traditional Employment Positions

Since AI has the potential to be a threat to society, AI is currently limited by society’s acceptance. There are plenty of AI substitutes for data entry office jobs, such as entry level accounting, finance, and human resources; however, there would be severe consequences to our economy if AI replaced the human workforce. For example, there would be a spike in unemployment rates until those workers learned the skills and knowledge necessary for a higher level job or found a job at the same skill and knowledge level that is not affected by AI.

According to the video ‘Humans Need Not Apply’, AI is capable of producing art, playing sports, being a doctor, and much more. The students who are currently studying the traditional post secondary topics will have the most problems when they graduate. Students pursuing degrees in humanities, English, math, biology, or chemistry will encounter the most challenges because their knowledge is not as adaptable. The cost of integrating AI into society no longer becomes an issue after decades of sunk costs into research but the economic and social factors of AI that would disrupt many industries. The video further explains the expendability of traditional employment positions.

A second limitation of AI is that it currently needs to be monitored by humans. By this restriction, most manufactures and service industries that would benefit from AI would rather integrate it into their current employees routines as a tool instead of using AI to replace the employee. This mentality limits the potential use of AI in society until it is sophisticated enough to make ethical decisions and better understand sentiment among many other things that make us human.

Ethical Implications

The possibility of machines that can think and learn raises a host of ethical issues. This section outlines the ethical issues of AI related to the transportation and health industries.

Transportation (Self-Driving Cars/Autonomous Vehicles)
The wide adoption of self-driving, autonomous vehicles promises to dramatically decrease the number of traffic accidents. Some accidents, though, will be inevitable. If an autonomous vehicle finds itself in a situation where it will hit a pedestrian or an oncoming car, what will it choose? How about if the vehicle has to choose between hitting a crowd of pedestrians and crashing itself into a wall, potentially injuring its occupants? Who is liable for this accident? A recent study called "Autonomous vehicles need experimental ethics," highlighted by The MIT Technology Review (2015) explores these questions, and more.

In terms of the legal side, it is not obvious if the passenger should be held legally accountable if the control algorithm ultimately makes the decision to hit a pedestrian, passer-by, or a wall. Some argue that liability has to shift from the passenger to the manufacturer since failure to anticipate these kinds of decisions may amount to negligence in design under the Product Liability law [1].

Since regulations lag behind technology in the case of autonomous vehicles, automakers of semi-autonomous vehicles are using workarounds. For example, to minimize manufacturer liability associated with their new ‘automated overtaking’ feature, Tesla Motors will require the driver to initiate the feature, thus ensuring legal responsibility for the maneuver consequences falls with the driver [2].

Defining the algorithms that will guide autonomous vehicles in situations of unavoidable harm is a difficult challenge, given that the decision is an ethical one. The control algorithms of autonomous vehicles will need to embed moral principles to guide their decisions in situations of unavoidable harm. However, if a manufacturer offers different versions of its moral algorithm, and a buyer knowingly chose one of them, is the buyer to blame for the harmful consequences of the algorithm’s decisions? Regulators and manufacturers will soon be in pressing need of answers to these questions, and that answers are most likely to come from surveys employing the protocols of experimental ethics.

Health Industry
The major issue with the use of AI in the medical field is determining where the responsibility lies when making potentially life-threatening medical decisions. For example, if a physician used IBM’s Watson to diagnose a patient with a rare illness and the diagnosis and treatment was deemed incorrect due to errors and the result ended in the patient dying, would the physician or Watson be responsible? Does the doctor have the final say when Watson suggests possible diagnoses so it would be the doctor’s responsibility to confirm all of the patient’s symptoms and concerns before administering a treatment, despite what Watson suggests? Alternatively, would the blame be placed on Watson for not understanding the symptoms or knowing that a given symptom for a given illness can manifest as a different illness? Also, would it even make sense to place the blame on AI as the repercussions or consequences would be insignificant, relative to the consequences a doctor could face. As the use of AI in medicine becomes more practiced, issues such as these will have to be addressed.

Future of AI

As AI continues to develop, the future remains clear. AI has the potential to improve the lives of its designers and mankind. Corporate ventures will create new venture projects to capitalize on AI advancements by introducing them to society to be accepted by the majority, just as it has been done for many other radical technological advancements. For future businesses to survive they will have to start adapting their strategies to be more forward thinking in order to keep up with the exponentially changing technology. With this in mind, society, not just businesses, will need to adapt and determine how to handle situations of AI that cause threats to society or ethical implications as it is integrated into society.

References

  1. http://www.brookings.edu/research/papers/2014/04/products-liability-driverless-cars-villasenor
  2. http://www.wsj.com/articles/tesla-electric-cars-soon-to-sport-autopilot-functions-such-as-passing-other-vehicles-1431532720
Personal tools