AI is an abbreviation that stands for Artificial Intelligence, which is translated as “artificial intelligence” in Japanese. The definition varies depending on the researcher, but it is generally used to refer to the science and technology of creating intelligent computer programs. We will explain the meaning, history, mechanism, and usage of AI.
What is AI (artificial intelligence)? Meaning and definition of words
AI is an abbreviation that stands for Artificial Intelligence, and is generally translated as “artificial intelligence.” The term was first used by computer scientist Professor John McCarthy in 1956 at a conference held at Dartmouth College in the United States.
Meaning of the word “AI”
AI is considered synonymous with “artificial intelligence” in Japanese. However, we need to be careful about what artificial intelligence means, as it varies depending on the situation and the person using it.
Professor McCarthy, who is considered an advocate of AI, defines artificial intelligence as “the science and technology of creating intelligent computer programs.” This definition is probably not unrelated to the fact that Professor McCarthy is a computer scientist. The definition of AI changes depending on the field of research and the topic surrounding it, so it is important to understand it in the context of the time.
Definition of AI by researchers
Japanese researchers also define AI independently according to their respective research fields. Here are some definitions.
Both are correct and appropriate definitions under certain circumstances. However, it may not be appropriate in other situations, so it is essential to understand exactly the context in which the terms AI and artificial intelligence are used.
History of AI (artificial intelligence)
It is said that research into AI has become active since the late 1950s, when the term AI was proposed by Professor McCarthy. We will briefly explain the history of AI until it became as familiar as it is today.
First AI boom (1950s-1970s)
The term “artificial intelligence” was first proposed at the Dartmouth Conference in 1956. In 1957, the concept of perceptron, which is the basis of neural networks, was devised, and development of perceptron began the following year in 1958.
For about 20 years after that, AI research became a boom among researchers. However, it became clear that perceptrons could only solve toy problems (problems in the world with fixed rules and goals, such as mazes or Othello), and the popularity of AI research gradually declined. This will usher in a period known as the “winter period” in AI research.
Second AI boom (1980s-1990s)
In the 1980s, technology called expert systems was created to accumulate knowledge in specific fields such as medicine and answer questions. Since this is a system in which knowledge is input in advance, it can be fully processed using one-to-one technology such as perceptron.
In 1986, a learning algorithm called backpropagation was devised. Backpropagation is a method that uses partial differentiation to calculate the path from the output layer neuron to the target neuron.
However, with the birth of deep neural networks, there will be limits to how the ability to express input and output can be improved by increasing the number of neurons.
In addition, in the 1990s, it became clear that expert systems were difficult to handle when the number of neurons became large, and that they were costly, and the popularity of AI research declined again.
Third AI boom (2000s-present)
In the 2000s, the practical application of “machine learning” in which AI itself acquires knowledge from large amounts of data has progressed. In 2006, “deep learning” was proposed, in which AI learns the elements (features) that define knowledge by itself, spurring the boom in machine learning.
In the 2020s, “image generation AI” that generates new images using machine learning and deep learning technology will be born. In addition, “ChatGPT” has appeared, which allows natural communication with people, automatic text generation, summarization, information collection, etc., and dialogue with machines has become a part of daily life.
:max_bytes(150000):strip_icc()/terms_a_artificial-intelligence-ai_asp-FINAL-ddba8ac599f3438d8064350d2ee1ae5a.jpg)
There are two types of AI (artificial intelligence)
AI can be roughly divided into two types: “Specialized Artificial Intelligence” and “General-purpose Artificial Intelligence.” However, as of 2023, general-purpose AI has not yet been completed, and specialized AI with limited applications is mainly used.
1.Specialized artificial intelligence
Specialized artificial intelligence is AI that specializes in a specific area. It can only be used for predetermined purposes, such as drawing pictures, recognizing words, and playing games.
In some fields, there are many things that have already been proven capable of surpassing humans. However, it is sometimes called “weak AI” because it has no applicability and the field is set in detail.
2.Artificial General Intelligence (AGI)
Artificial general intelligence is AI that solves a variety of problems in a wide variety of areas. It has self-control and self-understanding like humans, and has the ability to flexibly respond to situations that exceed expectations at the time of setup.
Unlike specialized artificial intelligence, it does not have a one-to-one correspondence, so it can also handle problems other than the programs and information input in the process of building artificial intelligence. Although it is still a fantasy as it has not been put into practical use, it is sometimes called “strong AI” in contrast to specialized artificial intelligence.
How AI (artificial intelligence) works
For AI to work effectively, two basic technologies are required: machine learning (ML) and deep learning (DL). Furthermore, when combined with rapid iterative processing of large amounts of data and intelligent
algorithms
, it can function as artificial intelligence.
We have summarized the terms used in AI mechanisms and related technologies.
machine learning
Machine learning is one of the basic technologies that support AI, and means that machines learn by themselves. Using machine learning technology, it is possible to self-learn and master even information and programs that were not input when building AI.
Machine learning is also used to process and analyze big data, which contains large amounts of complex data. Machine learning is an essential technology for responding to new information in a timely manner without frequent updates.
deep learning
Deep learning is a machine learning method that utilizes multilayered neural networks.
For example, unstructured data such as images and audio, which was considered difficult to digitize, is now within the scope of machine learning. Additionally, AI learning from unstructured data enables image and voice recognition, natural language processing, and more.
neural network
Neural networks are also a type of machine learning. It is an algorithm composed of interconnected processing units like neurons in the brain, and is characterized by a layered structure consisting of an input layer, intermediate layer, and output layer.
With a neural network, when a user presents a question and an example answer, the neural network can apply the given combinations to derive an answer that is close to the intended answer. As will be explained later, neural network mechanisms are also applied to natural conversational AI tools such as ChatGPT.
genetic algorithm
A genetic algorithm is an algorithm that derives the optimal solution using evolutionary methods, assuming that “excellent individuals” are “good answers.”
It is characterized by its ability to quickly determine optimal solutions for calculations that are difficult to do manually. In addition, even for problems that require testing a huge number of combinations to obtain an answer, it can arrive at the optimal answer in a short amount of time.
expert system
An expert system is an algorithm that defines rules based on interviews with specific experts about how to deal with specific situations, judgments, anticipated challenges, etc.
Since it is based on human experts, you can expect highly accurate answers to specific questions. For example, when a specific symptom is observed, an expert system equipped with a doctor’s judgment criteria can be used to instantly determine what kind of disease the patient may have.
cognitive computing
Cognitive computing is a subfield of AI and is a system that aims to establish natural dialogue between machines and humans. For ambiguous questions that don’t necessarily have one answer, we formulate a hypothesis and give the answer we think is the best.
It also has a deep learning function that stores feedback on the answers it derives and further deepens its learning. In contrast to AI, whose ultimate goal is to think on its own, cognitive computing can be said to be a system whose purpose is to support human decision-making.
computer vision
Computer vision is a system that performs image recognition using pattern recognition and deep learning. The recognized images are also processed as information and applied to foreign object detection, production management, etc.
Computer vision is the equivalent of human vision. It is widely used in manufacturing, energy, and automobile industries, and related markets are rapidly expanding.
Natural language processing (NLP)
Natural Language Processing (NLP) is a series of technologies that allow computers to process natural language that humans use on a daily basis.
It has a lot in common with cognitive computing in terms of technology related to dialogue. However, the difference is that while cognitive computing focuses on establishing dialogue, natural language processing focuses on understanding ambiguity and overlapping meanings.
Natural language processing analyzes the structure, syntax, and meaning of words in order to understand the language, and makes final adjustments by checking that they match the context. In addition to interactive search systems, natural language processing is used in a variety of situations that deal with words, such as character conversion prediction and translation tools.
Latest technology using AI (artificial intelligence)
AI is used for a variety of purposes.
In recent years, there has been an increase in the number of companies developing and providing conversational AI services such as “ChatGPT,” which allows commands such as calculations and searches to be performed through natural conversation. In addition, image generation AI, in which AI automatically generates images, and voice recognition AI, which recognizes human speech as speech and converts it into text, are also attracting attention.
Introducing our representative services.
Interactive AI chat service
Traditionally, when using search services such as Google and Bing, searchers had to enter potentially related keywords into the search window and browse the displayed web pages to find the answer themselves.
However, by using conversational AI, you can search for the information you want to know, or perform tasks such as calculations and table creation by inputting natural words similar to chat conversations. We will introduce some representative conversational AI services.
ChatGPT
ChatGPT is a service provided by OpenAI. With advanced AI technology, you can not only enjoy natural conversations just like
chatting
with a human, but also instruct the robot to create sentences, perform calculations, program, etc.
Since its release in November 2022, it has attracted attention around the world as an innovative service that can be used for free. In addition, many tools and services that utilize ChatGPT have been created.
It is expected that the use of conversational AI services will increase in the future, so it is necessary to understand the basics of how to use ChatGPT and what it can do. In the next article, we will specifically explain the basics and mechanism of how to use ChatGPT.
BingAI
BingAI is an interactive AI service from Microsoft, which has a large investment in OpenAI. By chatting with natural language, you can create sentences and images.
Additionally, Microsoft publishes Bing as
a search engine
, and BingAI allows you to check the content that can be searched with Bing, so you can use it not only to enjoy conversations but also as a search tool.
Although ChatGPT can also be used as a search tool, it should be noted that real-time information is not reflected, so there is an issue with the freshness of the information. If you want to use it as a search tool, BingAI would be more suitable as it has a search engine and also shows sources.
The following article explains in detail how to use BingAI and how it differs from ChatGPT.
Google Bird
Google Bard is an interactive AI that uses “LaMDA,” a large-scale language model developed by Google. As the name suggests, Bard uses a natural and rich vocabulary to perform tasks such as writing sentences.
Google Bard leverages Google’s strength in search functionality to provide real-time answers based on information on the Internet. When you want to look up related web pages, use Google Search, and when you want to get answers to your questions directly, use Google Bard.
Meta A.I.
Meta AI is an AI service that uses the large-scale language model “LLaMA” developed by Meta, and is also the name of the company’s AI research laboratory. In the future, CEO Mark Zuckerberg announced that SNS will be able to receive user support incorporating AI.
Meta also develops image segmentation services, voice translation services, and services for creating videos from text. Although there are many services that are limited to a limited number of users, many interesting services are expected to be made available to the public.
Image generation AI
Image generation AI is a technology in which AI automatically generates images by inputting text and data. The advent of Stable Diffusion, which can be used for free, has attracted increasing attention and is expected to be used in business.
In the next article, we have summarized image generation AI services that can be used for free. It also provides detailed explanations of what kind of illustrations are actually created and how to use them.
Music generation AI
Music generation AI is an AI service that creates songs and chords. It is used to create tracks from scratch in a fraction of the time, and to support the creative process of musicians.
Both use artificial intelligence algorithms that can generate new music by learning musical scores and audio data and creating patterns. There is also a method of learning music theory and sensibilities that can be reflected in new songs.
voice recognition AI
Speech recognition AI is an AI technology that recognizes the words spoken by people, converts them into characters, and converts them into text data. It is used for minutes creation, translation, call center customer service, etc.
Additionally, Google Assistant and Siri are search services equipped with voice recognition systems. Since you can skip the input process, it is useful for those who are not good at keyboard operations, or when you cannot take your hands off while working.
How to utilize AI (artificial intelligence) in each field
AI is used in various fields and industries. We will introduce fields and industries where artificial intelligence is already being actively used and how to utilize it.
manufacturing industry
The manufacturing industry has been using AI technology to promote automation. For example, we are introducing AI-equipped robots that perform simple and repetitive tasks to improve work efficiency and safety.
In addition, in inspection work, which is essential for improving a company’s reliability, AI systems such as computer vision are used to instantly determine whether a product meets the requirements and decide whether it passes or fails. It will be possible to check for minute paint misalignments that are difficult to notice with the human eye, and the inspection system will be significantly improved.
With the declining birthrate and young people leaving the manufacturing industry, the manufacturing industry is an area where securing human resources is expected to become increasingly difficult in the future. It can be said that it is necessary to utilize AI to improve efficiency.
medical field
In the medical field, the pursuit of accuracy and speed is an important element that can affect lives. AI technology has been used for image diagnosis, medical record analysis, and detecting abnormalities in patients, making it possible to make accurate and quick decisions and treatments.
For example, a doctor analyzes what kind of disease a patient may be suffering from based on test results and symptoms. In the past, patients relied only on the experience and knowledge of doctors, which often led to them not recognizing the possibility of a disease, resulting in treatment being too late.
However, if big data is used to analyze medical records, the possibility of illness can be detected early. Another advantage of using AI is that it is not influenced by the experience or knowledge of the doctor. By addressing the disease before it becomes serious, not only will patients be able to regain their health sooner, but it may also lead to lower healthcare costs for the entire country.
In addition to improving the efficiency of medical operations and the precision and quality of medical care, AI technology is being utilized to address the shortage of doctors and medical care in remote areas. The number of cases in which advanced treatment can be received regardless of location will increase, and medical opportunities are expected to become more homogenized and equalized.
Sports field
The field of sports is also an area where the use of AI is expected. For example, by utilizing and analyzing past data as big data, we are able to improve the quality of strategy formulation and training.
Additionally, image analysis from multiple angles using AI has made it possible to make more accurate decisions. Unlike refereeing by human eyes, it is difficult for assumptions or misperceptions to intervene, so it is also useful for avoiding trouble.
In addition, there is an increasing number of cases where spectators are provided with information that reflects player data in real time, and where dynamic pricing is introduced, which determines ticket prices according to the balance between supply and demand.
retail industry
In the retail industry, AI technology is increasingly being introduced to inventory management and ordering operations. Analyzing past sales data enables lean inventory management and helps avoid stockouts. Not only does it make it harder to miss sales opportunities, but it also makes it possible to sell without inventory, making it easier to start a business.
AI technology may also be utilized in store design. AI cameras analyze customer behavior in the store and change the store layout to make it easier to shop. This will make purchasing easier for users, and increase sales for sellers.
travel industry
Dynamic pricing using AI is also being used in the travel industry. By calculating package tour fees, accommodation costs, fares, etc. from the balance of demand and supply, you can set more reasonable prices.
Another advantage is that by offering low prices during periods of low demand, new demand can be generated and travel products can sell more easily. AI reduces labor costs, increasing the competitiveness of small and medium-sized travel agencies, and allowing them to increase sales based on their ideas.
What are the challenges facing AI (artificial intelligence)?
By repeating machine learning, AI becomes more versatile and reliable, making it possible to perform advanced tasks. However, it has been pointed out that AI learning data may be influenced by the developer’s personal attributes and values, which may be disadvantageous for attributes and cultures other than the developer.
Furthermore, as advanced image processing has become possible through AI, the issue of deepfakes, where celebrities’ faces are changed without their permission, has arisen. Images may have more spots or wrinkles than they actually are, or their appearance may be significantly altered in a malicious way.
Other problems have also arisen, such as AI being used for military purposes. As the image analysis ability of AI is increasing, it is thought that an increasing number of countries and regions are able to understand with high precision the bases and military facilities of hostile nations and groups. It would be fine if it acted as a deterrent against attacks on hostile nations or groups, but the problem is that it could lead to malicious use, such as targeting weakly defended facilities.
AI is a technology that has been developed through the concentration of human intelligence. In order to utilize AI in a way that is beneficial to humanity, both system developers and individual users need to have high ethical standards.
Utilize the ever-evolving AI (artificial intelligence) in your business
AI technology is evolving every day. Even those who are not comfortable with AI should already be using it through smartphones, credit cards, home appliances, etc. Rather than shying away from something because you’re not good at it, understand how it works and how to use it and actively use it to make your business more efficient.
However, when using AI technology, a high level of literacy and ethical standards are required on the part of the users. In particular, when handling personal information of customers or business partners, it is necessary not only to obtain permission from the other party, but also to be cautious in order to prevent leaks.
Furthermore, AI technology is not something that can be simply introduced. It is important to clarify what you want to achieve, define your objectives, and use them in the appropriate situations. Actively learn new technology, deepen your understanding, and apply it to your business.