INTRODUCTION
Artificial Intelligence is quite the buzzword these days. It is usually associated with words like Machine Learning, Deep Learning, Siri or Alexa, Self-Driving Cars, and on the extreme, Robots taking over the world.
I tested out building a simple Artificial Intelligence Project on Buildspace using GPT-3 text generation tool developed by Openai which generates natural-sounding text based on my inputs.
This prompted me to try out more use cases for artificial intelligence. The rest of this article was written with the aid of the GPT-3 text generation tool, enjoy.
What is Artificial Intelligence Anyways?
Artificial intelligence (AI), is a process of programming a computer to make decisions for itself, it mimics human intelligence. This can be done in several ways, but the most common is to use a set of rules to filter information and find recurring patterns. The computer then makes predictions or recommendations based on what it has learned from the recurring patterns.
There are generally 3 types of Artificial Intelligence:
Artificial Narrow Intelligence (ANI): This is the type of AI that is most commonly seen especially in consumer electronics and applications. It is designed to perform a single task or a limited range of tasks. Examples include digital assistants like Siri or Alexa, Self-driving cars, in web engines like Google to rank pages, videos, and the like.
Artificial General Intelligence (AGI): This is the type of AI that can learn and understand like a human. This AI is also able to reason and solve complex problems as a human would. Currently, there are no AGI applications in the market as this is still in development.
Artificial Superintelligence (ASI): This is the highest level of Artificial Intelligence often seen in science fiction movies. This ASI is capable of surpassing human intelligence and can be used for a variety of tasks. Currently it only theoretically exists as it has been shown that AI works best when programmed to achieve a single task.
Machine Learning
While Artificial Intelligence and Machine Learning are often used synonymously, they are not the same. Machine Learning is a subset of Artificial Intelligence, it is a tool/technology used in the field of Artificial Intelligence. Others include Natural Language Processing, Deep Learning, and Robotics.
Machine learning is a method of teaching computers to do things without being explicitly programmed. Basically, you give the computer a bunch of data, and it figures out how to do something with that data on its own.
For example, imagine you want to teach a child how to recognize animals. You could show the child pictures of animals and tell them what each one is called. The child would then learn to recognize animals by looking for certain features, like fur, four legs, and a tail. In contrast, with machine learning, you would give the computer a bunch of pictures of animals, and it would learn to recognize animals by looking for patterns in the data. The computer would not need to be told what each animal is; it would figure it out for itself.
Machine learning is powerful because it can automatically learn and improve from experience. For example, a machine learning algorithm might be able to automatically identify animals in pictures with 99% accuracy after seeing a few thousand examples. But if you showed it a million examples, it might be able to get 99.9% accuracy. The more data the algorithm sees, the more accurate it becomes.
Machine Learning and Data Science
Data Science is often mentioned in the discussion about machine learning and artificial intelligence. Data Science is extracting knowledge and insights from information.
Data Science is a field of study that involves using computers to analyze large amounts of data to find patterns, trends, and insights. It is a way of turning data into meaningful information that can be used to make decisions or solve problems.
Data Science can be used in a variety of industries, including healthcare, finance, retail, marketing, and education. For example, in healthcare, data science can be used to predict health outcomes and improve patient care. In finance, data science can be used to detect fraud, predict stock market movements, and develop investment strategies. In retail, data science can be used to personalize customer experiences, identify trends in customer behavior, and improve marketing strategies.
Realistic View of Artificial Intelligence
AI can be used to do many things, such as recognizing faces, driving cars, and understanding natural language. It can also help businesses make decisions by analyzing large amounts of data and creating predictive models. AI can also be used to automate mundane tasks, such as data entry or responding to customer inquiries.
However, AI is not without its limitations. AI cannot replace human creativity, intuition, and common sense. AI is also limited by the data it is given and its ability to interpret complex relationships. AI also faces challenges when dealing with unstructured data and faces ethical issues when used for decision-making. Ultimately, AI can be an effective tool, but its capabilities are still limited.
There have been talks about AI replacing skilled workers like Artists, Writers, and even Doctors. AI technology is used to automate certain processes and generate new ideas, but it cannot replace the creative and technical abilities of human artists. AI is being used to help doctors with data analysis, to predict potential health risks, and to inform treatment decisions however cannot provide human empathy and understanding. AI also cannot make decisions regarding complex medical issues that require human intuition.
Discrimination in AI
AI is used to automate several repetitive tasks, but its systems are designed and created by humans hence reflect the biases and prejudices of its creators. For example, facial recognition algorithms are often trained with datasets that are not diverse, causing the algorithm to have difficulty recognizing people of different races and genders.
AI systems that are used to identify people can be biased against certain racial or gender categories. AI systems can also be used to perpetuate existing inequalities, such as in the case of hiring algorithms that favor certain candidates over others. AI systems can also be used to make decisions about criminal justice, healthcare, and other areas of life that can have a disproportionate impact on certain groups of people.
Adversarial Attacks on AI Systems
Artificial Intelligence Systems are not fool-proof from hackers and bad guys. In 2019, researchers from the University of Virginia and the University of Illinois at Urbana-Champaign were able to hack an AI-based facial recognition system and manipulate it to identify people incorrectly. They used a technique called 'adversarial example generation' to generate images that the system misclassified. The researchers showed that such techniques could be used to bypass the security of AI-powered systems.
Adversarial example generation is a technique used in machine learning to generate inputs that will lead a model to produce an incorrect output. This is done by making slight modifications to an input that are imperceptible to humans but cause the machine learning model to misclassify the input. This technique is often used to test the robustness of a machine-learning model and ensure it is not prone to adversarial attacks.
For example, a machine learning model might be trained to recognize handwritten digits from 0 to 9. An adversarial example could be generated by making a small change to the image of a 9, such as adding a small line, that is unnoticeable to the human eye but causes the model to incorrectly classify the image as a 4.
Conclusion
Artificial Intelligence (AI) is the ability of a computer program or a machine to think and learn. It is a branch of computer science that deals with programming computers to do certain tasks that usually require human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages.
AI has the potential to revolutionize many aspects of our lives, from healthcare to transportation. AI systems have the potential to improve decision-making, increase efficiency, reduce costs, improve customer experience, increase safety, and reduce costs.
AI also has the potential to create new jobs and automate existing ones. As AI continues to progress, it is important to consider the ethical implications of using AI and to ensure responsible use and development.