Issue: Developing regulations concerning artificial intelligence
Forum: Special Conference 2
Position: President
Name: Fabianna Flores
Introduction
Over the years, there have been many technological advances linked to artificial intelligence. Computer scientists have the ultimate goal of getting computer systems to perform tasks that would usually require humans to use our intelligence. These tasks include those that require humans to carry out decision-making, perception, recognizing and translating speech. There have already been advances that have allowed models and computers to gather information and keep this information to be able to carry out tasks that need human intelligence. However, scientists have not yet had a breakthrough in getting computer systems to have the same complex thinking methods as seen in humans.
Furthermore, the term artificial intelligence was first introduced in 1956. Ever since then, scientists have continued to create models to further develop artificial intelligence. These include projects such as the Turing Test and IBM’s Watson. Both these models have used computer systems that are artificially intelligent as a means to show the development of the theory that computer systems can be as smart as humans, act similarly to humans and exceed human intelligence in some cases. The reason behind this is the fact that computer systems use algorithms in their neural networks so that patterns can be followed through inputs so as to produce an output using the information given. This reflects the way that humans would perceive the information and how humans would react to it, meaning that the output of the computer system closely resembles the conclusion that a human would draw from the situation. In the Turing Test, scientists were able to see this close resemblance between artificial intelligence and human intelligence.
Additionally, this has led to the growth of developing artificial intelligence since it would make tasks performed by humans more efficient and allow humans to use their skills elsewhere. However, problems do arise from the inevitable growth of artificial intelligence. The future of artificial intelligence shows the high possibility of high levels of unemployment, as computer systems and models will do many jobs. Also, the legal status of these systems will need to be considered when they begin to become more lifelike. Another possible problem is the fact that these systems may have negative effects on the world in terms of the environment. It is because of these possible problems that regulations need to be set up to prepare the future generations of the great impact artificial intelligence has and will have on the human race.
Definition of Key Terms
Algorithms
Algorithms are formulas used in computer programs that can solve a problem through assigning specific actions to the program. Therefore, a computer program will follow a set of rules (the algorithm) to calculate or solve a problem. In AI (artificial intelligence) the algorithm used tries to mimic the way the human mind would process the information.
Artificial intelligence
Artificial intelligence (AI) is the ability that a computer system has that enables them to mimic human intelligence. With the use of artificial intelligence, computer systems are able to do tasks that typically require humans to use their own intelligence to solve a problem or make a decision. These tasks include abilities such as making decisions, perceiving visual stimuli, recognizing speech.
Computer system
A computer system is a system that is able to execute actions or process information in an automated way. It uses a hardware system that works together with a software program that allows the computer to work and carry out a specific task.
Hardware system
The physical parts that make up a computer system such as the keyboard, monitor, and hard drive.
Human intelligence
The ability that humans have to use memory, imagination, reason and judgment to understand and make sense of their surroundings. It is used by humans to gather information to make decisions and solve problems.
Neural network
It is a computer system that aims to resemble the way the human mind and nervous system work. Neural networks are used in artificial intelligence since the end goal is to make a system that mimics the processes that occur in the human mind.
Software program
A software program is a program that does a certain task given to it through an algorithm. Examples of software programs include the Internet browsers Google and Firefox.
General Overview
Background Information
Artificial intelligence (AI) can be dated back to Greek mythology, when mechanical systems were shown to act with some form of intelligence. After World War II, humans were able to build computer programs to complete simple tasks and problems with the use of algorithms. However, AI was not formally introduced until 1956.
Developments
IBM, a computer hardware company, first introduced artificial intelligence at a conference in Dartmouth College in 1956. Here the theory of ‘artificial intelligence’ and its future was discussed between scientists. After this, many experts began developing systems that each had a specific task to complete which used AI. An example of one of these systems is the MYCIN system developed by Feigenbaim Bachanan and Edward Shortlife. The MYCIN uses an algorithm, which includes around 450 rules that it needs to follow in order to diagnose blood infections. This showed the possibility of using AI to aid humans with tasks so that they can be done more efficiently.
However, soon after the conference, the AI field went through a period of time known as the ‘AI winter’. It was called that way since not many advances were made, regardless of the amount of funding received for projects. This AI winter occurred from 1974 to 1981. It was at the end of this period that the British and Japanese regained interest. The British government began funding projects once again in order to compete with the advances seen in Japan. A couple of years later, in 1997, a breakthrough occurred when IBM created a computer system that was able to beat world champion chess player Gary Kasparov. More recently, in 2011, another artificially intelligent system was able to win the quiz show called Jeopardy!. Both these advances showed the possibility of using neural networks that can outsmart humans and respond in ways similarly to them.
Machine Learning
Many of the most recent developments have been based around machine learning, which aims to develop AI systems to respond to inputs in ways that resemble how humans would respond. An example of this firstly began in 1950. It is known as the Turing test, which aimed to test whether an AI system could be seen as intelligent by mimicking human conversation. Only one machine was able to pass the Turing test in 2014 when over half of the test subjects believed that they were talking to a human when in fact they had been talking to a machine. This success of the Turing test showed that AI systems have been developed to a point where they are able to mimic human conversations.
Many other developments in AI have been linked to transportation and healthcare. For example, Google has come together with UK’s National Health Service to develop a software that is being taught to be able to diagnose cancer from scans, identify heart disease and Alzheimer’s. This is linked to machine learning because these systems will be able to take in information, process it and act upon their software to improve healthcare.
The Future
The advancement of AI will bring many opportunities for humans. AI systems will be able to take over many tasks and therefore change our daily lives by making them more efficient. The main developments so far have been in developing machine learning, which is when a computer is able to improve its function by receiving more data and being able to process it faster to provide better solutions or recommendations. This has already been introduced to our daily lives through the use of technology that recognizes our speech and systems that recommend specific things to us such as movies. We have already begun to become reliant on these day to day uses of technology. It is because of that reliability that many companies will begin to use these wants of consumers to develop even more AI systems. The main goal of the future of AI is to improve education, healthcare and transportation.
Firstly, with the use of AI, transportation will be made easier through automated vehicles. These vehicles use artificial intelligence system to replace the driver, meaning an actual human driver is no longer required. This can already be seen through the fact that autopilot is an option in many vehicles. In 2012, Google tested a self-driving car, which was classified as one level below full automation by the US Department of Transportation. Other types of vehicles are close to becoming fully automated. This includes all trains being automated as this is already seen in the Vancouver SkyTrain. Also buses are close to being automated, with Mercedes-Benz has designed and tested an automated bus. The future implications of this are to make transport safer.
Another possibility that AI will bring to humans in the future is the fact that robots that use AI can perform dangerous jobs. These dangerous jobs include defusing bombs. Robots have already begun doing this but these robots require human control, as they do not yet use AI. Once further developments are made in AI, the machines will be able to work independently to prevent humans from harming themselves in dangerous jobs such as those that include toxic substances.
Additionally, artificially intelligent systems have the ability to retain more data than a human since they can store all the information given to them. This means AI systems could be given lots of data and information which can be processed into an output much more quickly than if a human did this. This process could potentially lead AI to use the information to make judgments and find solutions to large problems we currently face, such as climate change. This will benefit us, as they will be able to find patterns and trends that we may have never been able to notice. This has already been seen through IBM’s Watson AI system: it analyzed 25 million medical studies to be able to analyze 1000 oncology cases. In the results it was seen that 99% of the time this AI system was able to come up with the correct conclusion but what was surprising is that Watson was able to find something that these doctors had missed 30% of the time. This shows that in the future, AI will help humans in many different areas, because they have the ability to do similar tasks as humans can do, but more efficiently or even better.
Possible Threats
Artificial Intelligence has not yet reached its peak. Until now, we have only been developing AI systems that will potentially proliferate once more advances are made. Since this is a relatively new field, no regulations have been set up and the advancements are outpacing policies. This may cause even more problems, as there will be no way to regulate technology that has potential threats to society.
Potential problems include the fact that technology that uses AI will become more powerful and in the end may begin to be used for dangerous reasons. These dangerous reasons could include the use of robots that use AI to replace soldiers or also the creation of autonomous weapons. Therefore, a main factor that needs to be dealt with is cyber security that is linked to the development of AI.
Another problem is the fact that the revenue may be unfairly distributed in the economy. This unfair distribution will occur because currently there is a higher demand on a human workforce, but once AI advances, many jobs will be taken over by these systems. Once this occurs, a lot of the revenue will then be given to the companies that own the systems. Since many AI systems will take over jobs, the rate of unemployment will increase. There is also a fear that at one point humans will no longer be able to control AI systems. Therefore, steps need to be taken towards developing regulations to ensure that these threats are researched into and specific policies are made.
Major Parties Involved
United Kingdom
The United Kingdom has been contributing to the development of artificial intelligence through many of its start-ups and universities. However, the British government has not yet taken the impact that artificial intelligence could have on the future of the country into account. There could be a lot of opportunities but also many threats. It is expected that by 2035, the use of AI could add 814 billion USD to the country’s economy since it is likely that artificial intelligence will increase economic growth of many developed countries. Many Members of Parliament in the United Kingdom have noticed the absence of the government in educating and training people about AI. Since the 1980s, the United Kingdom has been in the lead in regulating technology, and therefore it would be wise for them to continue regulating AI so that AI benefits society and the threats are diminished.
United States of America
The United States of America has had extensive AI research seeing as a lot of funding is put into this development. Firstly, the Defence Advanced Research Projects Agency (DARPA) has funded projects that will later be used by the military. For example, the Mind’s Eye Program is a program being developed by the US military that will use an artificially intelligent camera so that the camera can make judgments and analyze what it sees. Also, many tech giants such as Google have been funding AI in the US. One of their main focuses is on speech recognition as seen with Apple’s Siri-feature. Google has stated that they are interested in robotics, particularly self-driving vehicles. Another source of funding is the National Science Foundation that fund projects that link neuroscience and AI. As we can see, the United States of America is very interested in the future of AI seeing that there has been a lot of funding for projects with different goals in mind.