Some of the common articles I see today are references to scientists either claiming AI will eradicate us all or that they disagree and will bring an era of change. However, before jumping down that rabbit hole, it may be better to understand what AI is because it is a very broad topic. The definition for AI or artificial intelligence is the following: “A branch of computer science dealing with the simulation of intelligent behavior in computers.”1Another popular topic that gets confused with AI is machine learning. Machine learning posses the following definition: “Machine learning is an application of artificial intelligence (AI) that provides systems the ability to automatically learn and improve from experience without being explicitly programmed.”2 In other words, machine learning is a topic within AI, but it’s by no means an interchangeable word. Machine learning is the principal of allowing computers to learn automatically by themselves. This can be confused as calling AI directly because it is technically replicating human behavior in terms of learning, but its main goals isn’t just replicating human behavior. Instead, its goal is to gather data and learn from it based on algorithms provided. These algorithms are usually handled by humans meaning that there is still human intervention with machine learning.
What is AI Then?
A better example of AI that people relate to doomsday terminators would be the sub field of robotics. Specifically, machines that replicate walking, jumping, running, talking, etc. are all great examples of the field of AI. This is a better example of replicating human behavior. The machine learning algorithm that Amazon uses to predict what you will buy this month isn’t really, but it is learning on its own which allows it to fall under the field of AI.
So, when we see articles about AI taking over the world or changing society, what are they really talking about? It’s mostly robotics. Going off the definitions, the terminator would be more of the doomsday AI, but a machine learning algorithm that can help figure out the political views of all of Facebook’s users and show them articles they would want to read in order to help persuade them that their political party is the right political party is a more realistic version of a doomsday AI. This kind of AI can trick people into gaining power dishonestly or they could maybe start a riot to help overthrow a key political position. What makes this category of AI so dangerous is that there’s no way to see who is controlling it. That type of investigation can take time and would be too slow to stop a crisis.
To keep it simple, AI is the equivalent to literature; the fields that fall underneath it such as machine learning and robotics are like the types of literature (fiction, non-fiction, etc.).
Why is the Topic AI so Big Today?
The topic of AI itself is such a big deal, because it’s beginning to change our personal lives. If you own Amazon’s Alexa and you use her to lock your door, order products, play songs, turn on/off the lights, then you are essentially allowing a form of artificial intelligence control a part of your life. Even if it may seem insignificant, AI itself is an extremely powerful tool. Machine learning has the potential to automate fields that used to be fully manual by humans. Now, the idea of hiring a full workforce to make predicate what customers are buying is silly; we can just leverage most of that with machine learning and have a few people manage it.
Then there are the ethical complications that come with the topic of AI as well. Just like automation, AI has the potential to be leveraged by people with ill intentions. Maybe the military wants to make predator drones decide for themselves where the target is a terrorist or civilian. Or maybe governments will try to control the general political opinions of it’s citizens by understanding what they like and manipulating how they react and feel about the news that occurs around them. These kinds of tools can be hijacked or misused to start wars or to turn elections illegally.
When we see top scientists say that AI is doomsday, they are looking farther into the future. One day the fields under AI will become advanced enough that they will not just predict what you will buy, but incorporate that into a shopping assistant who can speak to you and make recommendations. We may also have robotics advanced enough to replace human soldiers. These things won’t happen tomorrow, which is why some scientists today would tell you not to worry about it right now or even call it irresponsible to talk about AI doomsday scenarios. However, if we wait until the technology is available and being used in real life, then we may be too late in stopping something devastating from happening.