In our last post, I wrote about how AI revolutionized customer service by improving efficiency, automating routine tasks, and personalizing the customer experience. Then, I shifted gears and discussed the customer service landscape—how digital enterprises struggled to differentiate their products and services in a crowded marketplace where customers expect personalized content and will ditch brands who fail to deliver it.
My goal with the second blog post in this series is clarity. I’m going to provide a straightforward definition of what artificial intelligence actually is. Then, I will spend the remainder of this post explaining several other AI-related terms and concepts you’ve seen floating around online and in the media. Beyond basic definitions, I will provide concrete examples of real-world applications for each concept.
Now that I’ve set the stage, I will bring the technical jargon into the foreground. For all the hype around artificial intelligence and its implications for society, I’ve noticed that it’s still pretty hard to find a straightforward definition of what artificial intelligence is. I’ve also noticed that there are quite a few AI-related terms floating around out there that are confusing on their own.
Keep in mind that there are five more blog posts after this one. As I move forward and start talking in more direct terms about AI’s impact on customer service for online businesses, it will be essential to have as clear an understanding as possible of what these terms mean and the interplay between these different AI concepts and sub-disciplines, because these are concepts that I’m going to bring up repeatedly throughout the remainder of this series.
So, What Exactly is AI (Artificial Intelligence), Anyway?
Artificial Intelligence is a branch of computer science devoted to designing machines and computer programs that can simulate human intelligence. However, beyond this basic definition, things start to get complicated because there are different kinds of AI, and there’s also a whole constellation of terms and sub-disciplines that overlap with AI that business owners need to familiarize themselves with. So, with that in mind, I’ve put together this little glossary.
Chatbot
In many ways, chatbots have become the public face of AI in the twenty-first century. Chatbots are AI-powered software applications designed to interact with humans through verbal communication or typed commands. Chatbots are installed on websites to handle routine inquiries and perform simple tasks. For more complex issues, chatbots assist human beings by routing them to the right person in customer service or technical support.
The earliest chatbots were deployed on websites in the late 1990s. They could only handle basic inquiries and relied on a pre-formulated list of keywords and responses to interact with human beings. In the present day, chatbots have grown considerably more advanced. They can handle open-ended customer inquiries and complex applications such as hotel reservations and restaurant recommendations.
Deep Learning
Deep learning is another sub-discipline of computer science that overlaps with AI. Deep learning applications deploy multi-layered neural networks to generate high-level representations of the real world. Deep learning differs from machine learning in that deep learning doesn’t require structured data to learn and perform tasks. Deep learning applications are dynamic and can synthesize several kinds of data to perform a specific task.
Applications for deep learning include autonomous vehicles, which use deep learning algorithms to parse vast quantities of data to recognize objects in the physical environment, detect pedestrians, avoid other cars, and make all types of driving decisions. Deep learning applications are also popular with law enforcement agencies to extract evidence from photos, videos, and sound recordings.
General AI
General AI (sometimes called ‘strong AI’) refers to AI applications with human intelligence. General AI applications don’t exist yet, although many scientists and developers believe they’re not far off. General AI applications will be capable of performing any intellectual task that a human can perform, with the ability to comprehend, learn, and apply knowledge across various domains simultaneously.
Although there’s a degree of speculation in general AI, the implications of what general AI could do in the next few years are staggering. General AI has the potential to accelerate research in fields such as medicine or particle physics by analyzing vast amounts of data and automating many labor-intensive processes. General AI could also have a role to play in advancing other emerging technologies, such as quantum computing and nuclear fusion.
Generative AI
Generative AI is a sub-discipline of artificial intelligence centered on applications that can create new content based on text inputs or voice commands. Generative AI applications rely on deep learning models trained on massive datasets. These applications can spot patterns and other relationships within these datasets and then use this knowledge to generate entirely new content in text, images, or audio.
The most well-known generative AI application is OpenAI’s ChatGPT-4, which can generate detailed responses to user-supplied inquiries in direct, everyday language. ChatGPT-4 is a popular resource for everything from outlining to coding. Other well-known generative AI applications are Midjourney AI, which can generate detailed images based on textual inputs, and OpenAI’s Jukebox, which can generate original music by parsing massive datasets of songs.
Large Language Model
The term large language model refers to an advanced data architecture built to process and respond to human language and text inputs. Large language models are enormous and comprise billions of distinct parameters, enabling them to capture complex linguistic patterns. Large language models are integral for natural language processing, where they are used in text summarization and language translation.
Large language models can be found in various applications, such as content recommendations on streaming platforms like Netflix or Hulu, deployed to suggest personalized content to users based on past preferences. They are also increasingly popular for summarizing long articles or documents and code development, where they have been used to automate and debug code writing and assist in software development.
Machine Learning
Machine learning (which you’ll often see written as ‘ML’) refers to a branch of computer science that overlaps with AI. Machine learning involves applications that can analyze data to learn and improve at a given task without human assistance or explicit programming. Through the parsing of enormous datasets, machine learning algorithms can predict outcomes and make actionable decisions with no human input.
Machine learning applications are popular in fraud detection, where they are used to analyze patterns in financial transactions to identify suspicious behavior and prevent fraud, and medical diagnosis, where machine learning models can process patient data to diagnose diseases or predict health outcomes. Machine learning applications are also popular in commercial agriculture, where they are used to predict crop yields and identify diseased plants.
Narrow AI
Narrow AI (also called ‘weak AI’) refers to AI applications designed to perform simple tasks. Whereas general AI can perform complex tasks, narrow AI focuses on executing simple tasks without genuine consciousness or abstract reasoning. Narrow AI applications operate within a specific set of predefined parameters, as opposed to general AI, which can emulate human cognition across a broad range of activities.
Examples of narrow AI include virtual assistants such as Siri or Alexa, which are capable of performing simple tasks such as setting reminders, compiling shopping lists, or controlling other smart devices around your home, like a smart TV or a Nest thermostat. Other commonly used narrow AI applications include the user recommendation systems of streaming platforms like Netflix, which analyze user preferences to recommend new content.
Natural Language Processing
Natural Language Processing, (commonly referred to as NLP), is a sub-discipline of AI and computational linguistics focused on the interaction between computers and human language. NLP involves the deployment of algorithms and other models that allow computers and AI-powered applications to understand, interpret, and respond to everyday human speech, in a way that is both linguistically and contextually accurate.
NLP algorithms are commonly found in applications such as chatbots and other virtual assistants, which are required to understand and respond to human speech. NLP is also an integral component for applications such as text summarization, where it’s used to generate concise summaries of lengthy documents, and machine translation, where it’s used to automate translation and facilitate multilingual communication.
Neural Network
A neural network is a computational model inspired by the human brain. Comprised of interconnected nodes referred to as neurons, they are arranged in overlapping layers capable of processing and learning from vast amounts of data. Neural networks are used for advanced pattern recognition and object classification. Neural networks are integral to many AI applications, such as natural language processing and advanced autonomous systems.
In the present day, neural networks are deployed in various real-world applications. Neural networks power speech recognition in popular virtual assistants Siri and Alexa. They’re also commonly applied in healthcare settings, where they’re used in medical image analysis, disease diagnosis, and drug discovery. Neural networks are also used in gaming, where they’re used to develop character behavior and strategy.
Sentiment Analysis
Sentiment analysis (also called opinion mining) is a technology closely related to natural language processing. Whereas NLP is concerned with understanding and responding to human speech, sentiment analysis seeks to understand the emotional component of human speech through analyzing auditory, textual, and linguistic cues. Sentiment analysis applications classify textual or aural responses as positive, negative, or neutral.
Although sentiment analysis is a newer technology than NLP, the technology is becoming more widespread as AI continues to move into the mainstream. Sentiment analysis applications are becoming more prevalent as companies work to proactively manage their brand reputations and monitor customer sentiment on social media platforms. Sentiment analysis applications are also used for market research and customer feedback tracking.
All Definitions Aside, AI Technology has Actually Been Around for Quite Awhile.
I hope you’ve enjoyed this glossary that I put together. If you’ve made it this far, then by now, you should know what a neural network is and be aware of the subtle differences between machine learning and deep learning. In the next blog post of this series, I will talk about the history of artificial intelligence and how artificial intelligence has been a quiet presence in mainstream commercial applications for much longer than most people realize.
My primary goal with the upcoming third blog post in this series is to move you, my readers, away from viewing AI as an unforeseen, ‘black swan’ style disruption that will take away our jobs and destroy the planet. Instead, I plan to demonstrate that the advances in AI that have taken place over the last decade are the gradual culmination of technological advances that have taken place largely out of the public eye.
This Post Has 0 Comments