AI is a term that many students will have a general understanding of through entertainment and the media. It is important, therefore, to teach students about the reality of AI and its applications, to avoid any misconceptions.
What is AI?
The Cambridge Dictionary defines artificial intelligence as “the study of how to produce machines that have some of the qualities that the human mind has, such as the ability to appear to understand language, recognise pictures, solve problems, and learn”. It was the mathematician Alan Turing who first posed the question, “Can machines think?”, in 1950. He also proposed what would become known as the Turing Test, which states that a computer can be considered intelligent if it can convince someone that it is human by giving responses to ordinary questions (read the full description here).
Today, advances in AI have led to the development of systems and programs that help humans perform many routine tasks. Researchers are working towards machines that can learn as we do. These machine learning algorithms allow computers to create their own rules for solving problems, by “training” the AI with thousands of examples.
Broadly speaking, current AI technologies fall into two categories:
|Specialised or narrow AI||Specialised AI systems are designed to perform a specific task or work in a specific domain. Most of the AI systems that have been developed so far sit under this category. Examples include self-driving cars, computers that play chess, and human-language processors like Siri or Alexa.|
|General AI||A general AI system is defined by its ability to solve any generalised task given to it, much like humans. Many experts consider this to be the future of AI, though it is beyond our current capabilities. Google’s DeepMind is an example of a system that is being developed to try to produce super-intelligent general AI.|
Though we are far from the reality of a fully sentient machine, there has been much debate about the future growth of AI systems and the impact they will have on humanity.
The impacts of AI
Advances in AI throw up many ethical questions to consider. One of the most famous dilemmas asks whether the AI of a driverless car should be programmed to save pedestrians or passengers, should the car’s brakes fail. This is explored in MIT’s Moral Machine; this website is a nice way to introduce AI’s moral dilemma to students.
We need to decide who is accountable for the decisions AI makes, which could affect people in many other ways. For example, if a machine creates its own rules and mistakenly labels someone as a terrorist, that person would be affected emotionally, and potentially economically.
Another interesting ethical consideration is the question of whether AI can produce true works of creativity. Is there a difference between the algorithm AI uses and human creativity? Similarly, could AI be used to create large amounts of media to push an agenda?
Stephen Hawking and others famously signed an open letter to the scientific community, urging them to carefully consider the impact AI could have if not ethically managed. Organisations such as Cambridge University’s Centre for the Study of Existential Risk work on determining the near- and long-term security implications of AI.
The centre’s current focus is the risk of superintelligent AI, or machines that far outstrip human intelligence. Such AI could grant outrageous economic and military power to whomever owns it, or could make decisions with deadly unexpected consequences. Could these risks be mitigated before we developed such AI?
Examples of AI projects
The folk at Google are great believers in the power of AI to bring about positive societal change and improve people’s lives. These are some examples of Google AI projects (more can be found here):
Deep Blue and AlphaGo
These are both examples of machines developed to beat humans at board games. They form an important case study because of the rapid improvement in intelligence from Deep Blue to AlphaGo in a short space of time.
- Deep Blue was a computer developed by IBM to play chess. In 1997, it famously beat the chess champion, Garry Kasparov. It worked by analysing all the possible outcomes of a move in order to pick the best possible choice.
- AlphaGo is a computer program that plays the game Go. In 2015 it became the first computer to beat a professional human Go player. Unlike Deep Blue, which was programmed with moves, AlphaGo used machine learning to win.
Over to you
Earlier this week, you were given the idea of asking your students to summarise a topic in one sentence. Now it’s your turn. In the comments, sum up your thoughts on artificial intelligence in one sentence.