Artificial Intelligence has been studied for decades and is still one of the most elusive subjects in Computer Science. This is in part due to how comprehensive the subject is. AI ranges from complex machines to search algorithms used to play board games. A portion of people is left intimidated with all this, while some have no problem sharing the little knowledge that they have on the subject as loud as possible.
The first time that the term “Artificial Intelligence” was used was in 1956 by John McCarthy when he held the first academic conference on the subject.
Since then we’ve made great progress.
Machines can now transcribe a foreign language instantaneously, provide a piece of solid medical advice (it can detect cancer, for example), compose classical music albums, and even beat us at chess. With the use of Machine Learning many companies are saving millions and optimizing their work every year.
However, has this progress reached the stages we like to think it has? There are many misconceptions in this field and we’d like to explain a few of them.
- AI is way smarter than us.
This is one of the first things that pops into everyone’s minds when discussing future implications of AI.
In his book How We Learn: Why Brains Learn Better Than Any Machine… For Now, Stanislas Dehaene, a French author, and cognitive neuroscientist deals with theories about how human beings learn. Or, better said: why AI scientists are still failing to imitate our learning abilities. The author is also a professor of experimental cognitive psychology at the Collège de France, and the director of the Cognitive Neuroimaging Unit in Saclay. He explains the similarity between the human brain and machines: the thing is that the algorithms are designed to recognize patterns the same as the first stages of our vision. And that’s it. Because what happens next is something AI can’t deliver (yet) – the unconscious way our minds operate. The symbolic processing is taking over the stage and that is just a higher level of processing when comparing to the machines.
In other words, humans are simply able to get more from the information, by extracting a multi-layered meaning to it. So, the idea that a super-intelligent and conscious AI can somehow outsmart humans is pretty naive. The AI methods are still narrow, created to be applied to solve real-world problems.
- It will take all of our jobs.
The main fear is that it will devour a vast number of work opportunities, leaving humans with no need to do anything by themselves, but also without jobs. While new technology will always impact the division of labor, the predictions that go along with AI are not completely realistic.
While recent development focuses on creating practical and useful solutions in real-world problems, we can’t say any of the solutions can operate completely independently. The humankind still holds the key for any further technological advancement. It’s the human programmers, data administrators, and users that provide the necessary input for their learning and improvement. By lacking the basic key components of intelligence, AI and Machine Learning lack the ability of problem-solving and planning on their own.
What we can expect in the future is that:
The human factor will be involved in all activities it creates additional value in, if not – it will slowly lose its place.
This is not just due to AI, but more of a natural progression of things.
- Can AI be 100% objective?
Many people believe without a doubt that it can, however in many instances this is not the case.
Let’s not forget that people are the ones that create algorithms. They’re only as fair as the data that is feed to them.
There have been a few examples in the past and this is still something that proposes a challenge. One of the most known examples of biased, untrustworthy AI is the COMPAS system. This system is used in Florida and other states in the US. The COMPAS used a regression model to predict whether or not a perpetrator was likely to reoffend. Even though it was optimized for overall accuracy, the model still predicted double the number of false positives for African American ethnicities than for Caucasian ethnicities.
- AI is not creative.
This one is both a misconception and the truth.
Well, so far, AI technology created numerous unprecedented and valuable ideas. Of course, AI can’t do it on its own. But, when combined with human intelligence and intuition, it can design engines (like in case of Rolls Royce), for example. Its “autonomy” is put to good use also when it comes to music, pharmaceuticals, and various types of computer art.
- AI and Machine Learning are Compatible Terms
ML is actually a sub-field of AI, which is turn a sub-field of Commuter Science. The general definition that can be given for Machine learning states:
Systems that improve their performance in a given task with more and more experience or data.
Another term that is often mentioned in connection to ML is Data Science. This is a recent umbrella term that includes machine learning and statistics, certain aspects of computer science including algorithms, data storage, and web application development. Data science is also a practical discipline that requires understanding of the domain in which it is applied in. These solutions often involve at least a pinch of AI, but are not solely based on it.
- AI is a Brand New Science
Not even slightly. One of the first mentions of this concept was foreseen in the 1840s. An English mathematician and writer, Ada Lovelace, predicted some of it when explaining how the machine “might compose elaborate and scientific pieces of music of any degree of complexity or extent”, as she said in her own words.
In 1955, Newell and Simon designed The Logic Theorist, which is considered the first AI program. We’ve already mentioned that the person who coined the term Artificial Intelligence is John McCarthy. In 1956 he organized a conference “The Dartmouth summer research project on artificial intelligence” to draw the talent and expertise of others interested in machine intelligence for a month of brainstorming. And, so it all began.
When we look at AI now, we can see that we’re still far from developing general artificial intelligence. We’ve made significant progress, however we are allowing ourselves to get carried away by the same ideas and predictions that were made in the sixties and the eighties. Sensationalizations aside, we need to recognize that even though AI is not what the media presets it to be, it’s still a valuable asset that can help optimize and improve many fields.