Artificial Intelligence Explained


Artificial Intelligence (AI) involves using computers to do things that have traditionally required human intelligence. This means that AI makes use of algorithms to classify, analyze, and predict data. It also involves further modifying of the data by learning and improving over time. A common example of AI are of how search engines work; if you are searching for best remote desktop software, best restaurant to offer pasta around your area, or even where to purchase a particular item.

The question is: what is the difference between AI and regular programming? Regular programs strictly define all possible scenarios and only operate within those defined scenarios. AI, on the other hand, “trains” itself for a specific task and allows it to explore and improve on its own. We could call it good AI if the program is able to “figure out” what to do when it would face an unfamiliar scenario or situation. An example is that Microsoft Word can not improve its coding on its own accord, while facial recognition software can get better at recognizing as more data is “trained” for it to learn.

If you wish to apply or design an AI program, you require a lot of data. AI algorithms are trained using large datasets so that they can identify patterns, make predictions, and recommend actions, much like a person does.

Brief History of Artificial Intelligence

Image courtesy from University of Queensland Australia

Here is a brief history of AI, which is taken from The University of Queensland, Australia. Modern AI began in the 1950s with the view of solving complex mathematical problems and creating a ‘thinking machine’, the idea of which is now called Artificial Intelligence. There were two methods proposed; one used formal rules to manipulate symbols, which is a logic-based approach this is known as “good old-fashioned artificial intelligence” or GOFAI. The other approach considered how the brain works and focused on “artificial neural networks” which still need to be trained using certain methods to solve problems.

In the first 20 years of AI development, GOFAI was a more successful approach, which led to hype and funding from the government. In real-world settings, GOFAI did not achieved its projected outcome as predicted. Artificial neural networks as well struggled, in the 1970s the funding for such research dried up, leading to the research to slow down. In the 1980s, both AI methods got improved which made previously difficult problems more achievable, and AI seemed more promising once again.

Deep learning is one of the reasons why AI has spurred the interest of scientists since it is a type of biologically-inspired neural network, that harnesses the huge amount of data now available since the power and speed of today’s computational technology.

Having the benefit of enormous data sets, modern AI neural networks often exceed human performance in many tasks which includes pattern recognition and even playing games like chess, previously a very difficult goal for AI. These systems can learn from previous experience which is not possible for GOFAI.

AI’s progress might now appear like it’s not far from reaching human-level intelligence. AI still needs massive amounts of data to learn, unlike our brains, which can learn from a single experience.

Most scientists said that in order for AI to further advance, researchers needs to understand about the basic principles on how our brains function and the kind of biological shortcuts our brains take to complete actions

Modern applications of Artificial Intelligence

Common example of Artificial Intelligence in current use are: Machine Learning, Deep Learning, Neural Networks, and Evolutionary Algorithms.

Machine Learning

Machine learning algorithms identify patterns and/or predicts outcomes. Many large organizations use huge data sets related to customers, business operations, and the like. Human analysts have limited time and brainpower to process and analyze all of the data since the customer’s data is different and has many outcomes.

Machine learning can be used to predict outcomes in a given input data; one example of this is regression analysis, but on much larger scales with multiple variables. A great example of this is algorithmic trading, where the trading model must be to analyze vast amounts of input data and recommend profitable trades. Since the model is used in the real world, it can adapt and “improve” itself as time goes by.

Neural Networks and Deep Learning

Neural networks try to replicate the human brain’s approach to analyzing data;  it could identify, classify, and analyze diverse data, it could also deal with many variables, and find patterns that are too complex for human brains to recognize.

Deep learning is a subset of machine learning. If applied to a neural network, it allows the network to learn without human supervision from unstructured data. This would be perfect for processing the large volumes of that organizations collect. These could include but not limited to texts, images, video, and audio.

Neural networks are usually combined with deep learning and computer vision, this is the reason why people talk about deep neural networks which are a neural network with more than two layers, having multiple layers means having more analytical power.

Some real-world applications of neural networks are image recognition, which are used by computer scientists to identify common disease on agricultural farms by just using a set of pictures to recognize the pattern and characteristic of the disease, etc.

Evolutionary Algorithms

Evolutionary algorithms, which are a subset of machine learning, self-improve over time. This creates a population of algorithms and preserves the ones most successful at predicting outcomes. This follows the “survival of the fittest” principle, the best outcomes and predictions are kept alive and the bad predictions are discarded. Some parts of the code from the winning algorithms are used to create new algorithms, and the selection process repeats.

This subset of machine learning is well suited to optimization tasks where there are a lot of variables and a dynamic environment.

Examples of powerful Artificial Intelligence in use today

Artificial Intelligence is commonly seen in households that contain smart devices or “smart technology”.

Automation: repetitive back-office tasks such as invoicing, and management reporting can be automated to save time and improve efficiency.

Customer service: Chat bots are one best examples of this; if you require support you are most likely talking to a chat bot first before reaching a human agent.

Social Media Platforms: Facebook uses AI to recognize faces; you could see this if you upload images on Facebook and it predicts the person’s face within your friends’ list.


Siri is a popular personal assistant offered on Apple devices especially for iPhone and iPad. It is a female voice-activated assistant that users interact with, to perform certain tasks or assists with find some information.

Siri uses a machine-learning technology in order to get smarter and capable to understand language on different questions and requests.


Netflix – a widely popular content-on-demand service uses artificial intelligence to process predictive technology, to offer recommendations on the basis of consumer’s interests, choices, and behavior by examining the number of records that the user watched or the similar genre that the user is into.

One disadvantage of this technology used by Netflix is that small or indie movies go unnoticed while big films continue to grow and are noticed on the platform hence this is an issue that needs to be remedied in the future.


Echo is a smart AI device available from Amazon, which is constantly becoming smart and adding new features regularly. It is a smart product that can search the web for information, shop, control lights, switches, thermostats, answer basic questions, read audiobooks, get information about the traffic and weather by using the Alexa Voice service.


Automobiles are also taking advantage of artificial intelligence trends, with Tesla leading the automotive market in this regard. Tesla’s cars have been able to provide many features such as self-driving, predictive capabilities, and even self-parking. Tesla cars even get update themselves to further improve their features.

Lyft and Uber

Uber and Lyft are popular ridesharing apps; with artificial intelligence, these apps could determine the price of your ride, minimize the wait time of the car and optimize the route to match with your partner along the way. With Artificial Learning, these apps are able to also calculate the demand having a ‘surge pricing’, ETAs of your pick-up point and destination, as well as delivery times for apps like UberEats.

Apple Pay

Mobile payments such as Apple Pay use artificial intelligence as well, in order to buy music, movies, and apps. With Apple Pay, users could buy nearly any product by simply holding an iPhone over an NFC (Near-Field Communication) terminal at the store. This lets users pay others and maintain funds using its Apple Cash feature. Fraud detection is one of Apple Pay’s Fraud Protection uses artificial intelligence in order to monitor and learn from your spending habits when using Apple pay features.


The concept of artificial intelligence is not a new one, and the technology continues to improve and influence the way we live, interact, and make our lives easier and better. There is much more to expect in the coming years with more improvements, development, and governance.

4 Replies to “Artificial Intelligence Explained”

  1. Interesting. What shall start off with to have my own simple AI system?
    I know Python is widely used to code machine learning systems. What do I need? Any guides?

  2. I have seen many tutorials around the web that gives you basic machine learning systems, one thing I tried before was the dogs vs cats where the computer predicts if it is a dog or cat with the test picture and I had fund with that.

Leave a Reply

Your email address will not be published. Required fields are marked *