Have you ever wondered what AI is?
Not a week goes by that we don’t hear about major advancements in artificial intelligence (AI) or how robots are outperforming humans. But many of us hear the term “AI” and silently wonder, “What is it?”. So let’s go a bit deeper into the rabbit hole that is AI so that all of us can hitch a ride on the future train.
What is artificial intelligence (AI)?
Artificial intelligence is an area of computer science that aims to create machines that react, work and even think like humans do.
Research into artificial intelligence focuses on four main parts. The first, and core part, is knowledge engineering. The idea is that machines can act like humans provided that they have access to an abundant amount of information relating to the world. The second part is that of machine learning. It is about giving machines the ability to learn without having been explicitly programmed to do so and essentially teaching it to think for itself.
Next, we have machine perception. This relates to a machine’s capability of interpreting data in a similar way to how humans use their senses (like sight) to interpret the world around them. Lastly, there is robotics. In this field, engineers aim to make it possible for machines to manipulate objects and navigate the physical world around them. An example of this is where a robot picks up a box and moves it to a different location.
What types of AI exist?
* Weak AI (or narrow intelligence)
This is a non-sentient (not having the ability to perceive or feel things) AI that only focuses on a narrow task or has the ability to only solve a very specific problem.
Apple’s Siri is a weak AI. It has a limited range and cannot move outside of certain boundaries. For example, there is a list of questions you can ask Siri that she will understand and respond to but once you ask a question not on the list, she cannot adapt nor answer appropriately.
* Artificial General Intelligence (or AGI)
A machine that possesses AGI can successfully perform any intellectual task that a human being can. The requirements for a machine to be classified as having AGI include being able to plan, learn, communicate in natural language, represent knowledge, reason even in uncertain circumstances and so forth.
Scientists are still trying to come to a consensus as to what tests a machine needs to pass in order to be classified as AGI and the following have been suggested:
– The Coffee Test
The machine has to go into a standard home and figure out how to make a perfect cup of coffee. This seems pretty simple on the surface but there are a lot of small, complex actions that need to be completed to make a cup of java that we, as humans, take for granted.
– The Robot College Student Test
A machine is given the task of enrolling in a college, taking and passing the same classes as a human would have to, and eventually getting a degree.
– The Employment Test
The machine must work an economically important job and be able to do the job at the same level or better than a human.
This is currently a hypothetical stage. Proponents of this theory posit that there will be a machine whose intelligence can ultimately surpass even the most intelligent and brightest human minds. However, there are differing opinions as to whether superintelligence is even possible and whether machines will ever be able to possess intentionality and first-person consciousness. Remember, the creators of robots are humans and humans themselves have intellectual limits which in turn limit the potential of AI.
Where is AI currently being used?
Many of the things we currently use were originally developed in an AI environment to solve problems in computer science. These include the computer mouse, automatic storage management, graphical user interfaces and object-orientated programming (like Java, C++ and more). They are no longer considered to be part of the AI field because as John McCarthy (one of the founding pioneers of AI) noted “as soon as technology works, no one calls it AI anymore”.
Banks use artificial neural networks to detect changes or claims that are outside of the norm which are then flagged for subsequent human investigation. In hospitals, these same networks assist in providing more accurate medical diagnosis. Furthermore, the medical field also employs AI for use in companion care robots for elderly patients, heart sound analysis and the interpretation of complicated medical images.
Robots are increasingly being used to do automated jobs in heavy industry. They are especially useful for jobs that are too dangerous for humans to do or repetitive, tedious jobs (this helps avoid mistakes that can happen due to human lapses in concentration).
The toy and gaming industry is probably the most visible in terms of the use of AI. The Tamagotchi, iPod Touch and Furby are all examples of AI-based inventions. If you play a lot of video games you might be familiar with video game bots. In a first-person shooter game these bots are able to mimic how humans play the game and you’ll often experience them in gameplay where you’re pitted against the computer. There are also NPCs or non-player characters. These are characters controlled solely by AI and not playable by humans. Think about MMORPGs (massively multiplayer online role-playing games) like World of Warcraft where you have to hand in quests to certain characters – these are NPCs.
AI tools are also being deployed for use in homeland security, data mining, e-mail spam filtering and various types of recognition (object, speech, text, voice, facial and gesture). There are also virtual personal assistants (like Siri), smart cars and smart homes that are becoming an ever-increasing part of daily life.
In next week’s installment of this two-part series, we’ll dive into the controversies that surround AI development as well as what lies ahead for the future of AI.