Artificial Intelligence or AI has been receiving a lot of publicity over the last few years. When AI beat the best Go players, including the world champion, people started taking it seriously. When it defeated a top-ranked Dota 2 player, it gained even more publicity. Then came autonomous cars, and people really went crazy over AI. Everyone began calling it the electricity of the 21st Century and the most important technological advancement of our time. On the other deal, some saw it as a threat, as something that could virtually replace humans as the apex predator.
Myths and Facts About AI
There are a lot of tales and truths about artificial intelligence that have been circulating for some time now. The idea of AI is not exactly new. It has been the subject of many science fiction films like Terminator and 2001: A Space Odyssey. Today, however, thanks to more powerful and accessible technology, AI is a good thing. But the term itself is so nebulous it is hard to define AI very clearly. There is also a lot of fanfare surrounding the technology, making it difficult to separate fact from fiction.
Can an AI order pizza for me from my Spectrum home phone? Can it create art? Can it bring about the destruction of the human race? These are questions, almost everybody with even the most basic understanding of artificial intelligence asks. But too many people know too little about the realities surrounding this new technology, being influenced by hyperbole and science fiction movies. This blog explores common questions about artificial intelligence and the myths and facts surrounding them, including:
- Is AI a Clearly Defined Field of Study?
- Can a Single AI Solve All Problems?
- Does AI Make Human Intelligence Obsolete?
- Does AI Need to Learn Constantly?
- Can AI Be a Threat To Human Existence?
Let’s take a closer look below.
Is AI a Clearly Defined Field of Study?
Artificial intelligence is a term that only dates back to about 60 years. Over these years, excitement surrounding AI has risen and dropped. A lot of what we know about AI is what the media has fed us in layman terms. Overly simplifying the term has led people to believe it is a distinct and precise field of study.
The reality is, there is not a lot of consensus on the term “intelligence”. As a result of this, researchers and scientists studying AI have pursued diverse fields. Some have studied pattern recognition; others have focused on problem-solving, while still others have looked at decision making. The result is that AI is not a unified field of study. It is a field with many inter-disciplinary areas and smaller fields within larger ones.
Can a Single AI Solve All Problems?
Many people, as a result of the hyperbole in the first section, believe that a single AI can solve all manner of problems, especially in the business world. This is a myth. The fact of the matter is that people don’t understand the subfields in AI and what their limitations are. People also tend to underestimate the human cognitive ability too much.
For instance, an AI can help a business manage its digital assets so that employees can have authorized access to any asset they need at any point. For a large enterprise, this would be a daunting task for human workers, and AI has definitely made life easier. But ultimately, the parameters within which the AI manages digital assets are still designed and tested by humans. Human skills are still needed to correctly create a system, measure its performance, and interpret its results.
And that’s just one aspect of modern business. There is no single off-the-shelf AI solution as of yet that can solve all business problems in any use case. A single AI solution is not likely to match, let alone surpass, human activity and intelligence in business contexts at this time.
Does AI Make Human Intelligence Obsolete?
There is a common myth surrounding AI that it will make human intelligence obsolete. The obvious fallout from this would be AI taking over human jobs, pushing people out of work and into poverty. Many data scientists claim they only need data to solve problems, not any knowledge of the domain itself. This may be true to some extent. But humans are the ones framing problems, creating hypotheses, and evaluating results.
The fact is that AI and human intelligence can both augment each other’s capabilities. There will always be a spot for human ingenuity, even in an advanced AI world. Humans and AI working together can accomplish far more than either could alone. That means artificial intelligence may not be replacing human intelligence any time soon.
Does AI Need to Learn Constantly?
Many people believe that the outcome of an AI depends on how you model it. The reality is that you can’t always be certain of the outcome. You will need to train your AI on data, and this training data needs to be an accurate sample of real-world data. Otherwise, the AI might work fine in testing, but the outcome will degrade when you put it in production. AI training is never complete in the real world. Depending on how fast real-world data changes, your accuracy will begin to decline. At that point, you will need to retrain the AI on new data sets that represent the changing situation.
Can AI Be a Threat To Human Existence?
This is something that many leading figures like Bill Gates, Elon Musk, and Stephen King have voiced their belief in. The theory is that an advanced AI system might be able to create better and more sophisticated versions of itself. This would let it quickly surpass human intelligence and adopt its own values different from the ones it was programmed with. What happens then is anyone’s guess, but most people believe that AI could very well decide humanity needs to be eliminated for the better.
To many, that seems the obvious conclusion that an emotionless, completely logical intelligence would arrive at when considering solutions to the biggest problems the world has. Human beings are the cause of war, poverty, global hunger, death, disease, and pollution. We have been harming the very world we live on ever since the first Industrial Revolution. It would make perfect sense to our hypothetical AI to decide that for the planet, and it’s billions of species to survive, humankind has to die. This is a theme echoed in many science-fiction movies like the Terminator saga.
Of course, the problem with this theory is that superintelligence is not yet a technical reality. Most of modern AI fails to approach even general intelligence. At best, it can deal with particular problems in a defined framework. AI is still in its infancy. That means you still have a lot of time to enjoy movies on Charter cable TV before machines take over the planet. But even we can’t guarantee that will never happen. In the meantime, let’s focus on using AI to solve real-world problems and make the world a better place.