Is AI really equipped to takeover the world?
It’s high tide for spooky season. People everywhere are watching horror films and going to haunted houses just to scare themselves for entertainment. Controlled creeps can be fun, but things that are out of our control can be genuinely scary. The uncertainty of the future, for example, can be nightmare-inducing for some. For others, something like AI will make their skin crawl. The thought of AI doing something it’s not designed to do to can be a spooky concept (we’ve all seen how Terminator plays out, right?) but how likely is that to happen in the real world?
Whatever your feelings are about spooky season, fact is, AI is here to stay and there’s nothing to worry about. We’re debunking some of the most popular myths that make people feel apprehensive about AI so you can sleep easy knowing there's no Cylon plotting to destroy the world.
Movies are Entertainment Not Reality
We look to film for entertainment, but sometimes a film just hits too close to home--ever feel like you're watching your own life, dreams or fears on a screen? Think about The Terminator, Blade Runner, 2001: A Space Odyssey, and The Matrix. What do all these films have in common? They portray a futuristic civilization where AI has taken over control.
A good portion of AI myths that have been exaggerated and exacerbated by movies:
- AI taking over the world
- AI ruling humans
- AI becoming sentient and thinking like humans
- AI becoming so advanced that humans can no longer control it
In these films, the world is often a dystopian society. The city is in ruins, machines walk about freely while humans cower in the distance. The storyline contains a common theme; these AI systems and robots were created to help humans and, at some moment in time, grew sentient and felt they were treated unfairly. These robots band together and decide to fight back by reshaping the world in their image, leaving only a few determined crusaders to save the world for the human race. Though this small band of misfits is oft to win the day and reclaim the world for humans, a seed of discomfort of a near future has been planted for some viewers.
Good thing, movies are just great at make-believe. First and foremost, it’s good to take a step back and remember that we are just in the early stages of AI. Human thought is vastly intricate, and while some may view AI powered programs and robots as being capable of their own thoughts, it’s actually incorrect.
Don’t Worry, AI Can’t Takeover the World: It's Just Not Smart Enough
Let’s debunk this. The data lifecycle of training an AI algorithm goes like this:
- Obtain training data based on what you need the end application to do.
- Annotate that data.
- Feed it into the model.
- Confirm that model is functioning as intended.
- Start again.
The great thing about machine learning is, once you give it enough data, it can identify and interact with new data that's a variation of the previous data it was trained from. If something occurs that the model was never trained to account for, then it doesn't take an action . For example, if you trained an AI model to identify food images in photos, it won’t magically be able to compose a symphony if you asked it to. Extending this to the concept of an AI robot taking over the world, You would have to provide the right training data to cover every possible scenario for it to learn to take over the world. Since no one wants that, AI will not be trained to do it.
Furthermore, human brains don't have a uniform way of thinking. Robots, however, do. Machines are trained how to think (or in actuality how to operate) by humans. They are incapable of thinking of new ideas--they simply lack the training data in diverse categories to develop sentience or human-level intelligence. One might argue that Terminator robots successfully took over because they were trained as military machines. While that’s true, don’t forget how difficult it was for the Terminator to finding Sarah Connor and killed any person with the same name. The AI wasn't programmed with the skills or enough data to correctly identify the right person on the first try. A human wouldn't make the same mistake. Movie robots have flaws and limitations just like robots of the real world do. The Terminator couldn’t exactly get his one job right, if he had, the movie would have lasted all of 5 minutes.
Movies are great for make-believe but, to be safe, Elon Musk created a company called Neuralink specifically designed to prevent any kind of Terminater / Blade Runner-esque takeover.
This spooky holiday season, rest easy knowing that movies are just movies with amazing special effects--many done with the use of AI, you can read our article May the 4th Be with You: The AI Behind Movie Magic to learn exactly how it’s done. AI is a tool to enhance and make our lives easier, so you do more and enjoy more, if you let it.
Stay tuned as we debunk more myths showing you why AI is great!