As artificial intelligence (AI) spreads throughout society today, the adoption and use of AI has a significant impact on the competitiveness of companies. For example, from manufacturing to healthcare to marketing, AI’s improved operational efficiency and advanced data analysis is one of the backdrops for major changes in our lives. Therefore, as AI technology evolves, companies are required to respond to improve business performance and realize innovation.
However, there are also many challenges and risks involved in implementing and utilizing AI. It is true that there are many points to be aware of, such as implementation costs, data dependencies, and security issues.
Therefore, in this article, we will explain what artificial intelligence is in the first place and its meaning in an easy-to-understand manner, including its types and what it can do. We will also introduce how the world will change with the development of artificial intelligence, so please refer to this article.
What is Artificial Intelligence (AI)?
First, we will discuss the three fundamentals of artificial intelligence (AI) in turn.
Definition of Artificial Intelligence (AI)
Every day, our brains process a great deal of information received from the outside world, making judgments and inferences. The technology to reproduce such human intelligence by computer is called Artificial Intelligence (AI).
At present, however, there is no strict definition of artificial intelligence.
This is because there are some gaps in interpretation and recognition among researchers and research institutions, and the definition is not unified.
Many people think of artificial intelligence as a robot that can think of anything by itself, like “Doraemon. This is called “general-purpose AI,” but in reality there are no examples of its practical application in the world yet.
Therefore, most of the time when we generally speak of artificial intelligence, we are referring to “specialized AI” that can only perform specific processing such as face recognition.
How Artificial Intelligence (AI) Works
Since artificial intelligence is a computer technology, it is basically driven by programs (instructions to computers). Programs embedded in artificial intelligence process large amounts of data and operate as if a human being were making decisions and inferences.
There are many types of artificial intelligence, such as image recognition and chatbots, and there are various ways to feed data to artificial intelligence.
For example, in the case of “self-driving cars,” sensors installed in the vehicle acquire information about the car in front and other information, which is converted into data and processed by the artificial intelligence.
History of Artificial Intelligence (AI)
Artificial Intelligence (AI) first became widely recognized at the Dartmouth Conference in 1956. Dartmouth professor John McCarthy’s use of the term “artificial intelligence” popularized AI among scientists.
Here is a brief history of artificial intelligence, which has seen four booms.
- Simple rule-based artificial intelligence (AI)
- The emergence of expert systems.
- Machine learning and deep learning (deep learning).
- Emergence of ChatGPT: Reinforcement Learning and Generative Models
Simple rule-based artificial intelligence (AI)
The 1960s, when the rise of artificial intelligence began, was a time when simple rule-based AI emerged and attracted much attention. AI in this era was strong in puzzle problems such as chess and Sudoku, and demonstrated a certain level of ability within a set set of conditions and rules.
However, the first boom in artificial intelligence did not last long. The reason for this is that simple rule-based AI had a weakness that prevented it from being applied to solving complex problems in the real world.
Therefore, people at the time eventually lost interest, disappointed by AI’s success in simple games and test cases but its inability to handle more complex challenges.
The emergence of expert systems.
Expert systems, which suddenly appeared in the 1980s, opened a new page in the history of artificial intelligence. Expert systems are high-level technologies that aim to solve more advanced problems by providing computers with a certain level of specialized knowledge.
Specifically, they were used in medical diagnosis and stock price prediction, and proved effective in situations where expert-level knowledge was required.
However, like the first boom, the second boom did not last long. The main reason was that the data collection and processing required a great deal of resources. The process of incorporating expert knowledge into the system was cumbersome and labor intensive, and the costs associated with the work process skyrocketed.
As a result, these cost challenges could not be overcome, and the boom gradually began to subside.
Machine learning and deep learning (deep learning)
Since the 2000s, machine learning and its branch, deep learning, have gained momentum in the field of artificial intelligence. In particular, the dramatic advancements in deep learning technology have brought artificial intelligence into a new phase.
Specifically, self-driving technology for automobiles has made significant progress. Tesla’s “Autopilot” and Google’s “Waymo” self-driving cabs are typical examples.
In image recognition technology, X-ray image analysis in the medical field and facial recognition functions in smartphones have been put to practical use. Natural language processing has also evolved, and the improved accuracy of Google’s search engine and multilingual translation services is another result of deep learning technology.
The third boom, which has had a direct impact on modern life, has spread to a wide range of industries. It has produced tangible results in each area and changed our lives in ways that make them more convenient.
Thus, thanks to advances in machine learning and deep learning, a world that was once thought of only in science fiction is now a reality.
Emergence of ChatGPT: Reinforcement Learning and Generative Models
In recent years, much attention has been paid to the evolution of reinforcement learning, including the emergence of ChatGPT and other generative models. The fourth AI boom brought about by these technological advances conveys the fact that we have entered an era in which artificial intelligence is capable of addressing a greater variety of problems than ever before.
Generative models such as ChatGPT have made operations more efficient in many industries and domains, from customer service to digital content creation to programming code generation. In particular, this type of AI is characterized by its ability to understand and generate natural language, and in an increasing number of cases, it is capable of conversing at almost the same level as humans.
Meanwhile, in the field of reinforcement learning, applications are expanding in a wide range of areas, such as AlphaGo, which appeared in 2016 and outperformed human professional Go players, energy optimization, and quality improvement in the manufacturing industry.
Through these advances, artificial intelligence has evolved from a mere “support tool” to a potentially useful tool for strategic decision-making and creative work, as evidenced by the worldwide proliferation of ChatGPT, the potential of this fourth boom in particular is immeasurable, The potential of this fourth boom in particular is immeasurable, and its future development is highly anticipated.
Main types of Artificial Intelligence (AI)
The following is a summary of the two main types of Artificial Intelligence (AI).
- General-purpose AI and specialized AI
- Weak AI and strong AI
General-purpose AI and specialized AI
The first classification is “general-purpose AI and specialized AI”. The differences between each are as follows
| General-purpose AI | Specialized AI | |
| What you can do. | Theoretically, perform a wide variety of tasks in one hand. | Specialize in specific tasks |
| Specific Examples of Utilization | Conversation Math problem solving Cooking recipe creation Home appliance control | Image Recognition Voice Recognition Automatic driving Weather forecast |
| technical limitations | Computational resources are very expensive Insufficient data Development requires a lot of time and effort | Tasks are limited but highly optimized |
| In the real world Utilization in the real world | Not in practical use at this time. | Widely used (smartphones, automobiles, etc.) |
| promise | Requires long time and high resources, but expected to be multifaceted | Tasks are limited but highly effective within their scope |
First, general-purpose AI can theoretically perform a wide variety of tasks in a single hand, just like humans. For example, in the future, it is expected to have not only conversational skills, but also multifaceted capabilities, such as solving mathematical problems, creating recipes for cooking, and controlling household appliances.
However, due to technical limitations and lack of computing power and data, general-purpose AI has not been put to practical use at this time. Computational resources are very expensive and development requires a great deal of time and effort.
Specialized AI, on the other hand, is AI that functions specifically for a particular task. Typical examples of specialized AI applications include image and voice recognition, automated driving systems, and weather prediction systems.
For example, specialized AI-based facial recognition systems are often used in smartphones and security gates. Other applications include weather forecasting, which analyzes large amounts of weather data to provide more accurate forecasts.
As described above, general-purpose and specialized AIs are classifications that focus on what tasks they can perform.
Weak AI and Strong AI
The second classification is “strong and weak AIs”. The differences between each are as follows.
| Strong AI | Weak AI | |
| What you can do. | Theoretical, emotional, self-awareness, ability to learn | Specialize in specific tasks or problem solving |
| Specific Examples of Utilization | Advanced humanoid robots in science fiction movies and novels | Google Search Siri Amazon Alexa, etc. |
| human nature | Theoretically, almost indistinguishable from humans | No ability to do anything other than what is set up. |
| In the real world Utilization in the real world | Practical application in the real world is still a long way off (the world of science fiction movies) | Widely used |
| Learning thinking ability | Theoretically capable of thinking, learning and understanding like humans | Advanced learning and solution skills for specific tasks. |
| Emotions self-awareness | Theoretically, they could have emotions and a sense of self like humans. | It is only a machine, with no feelings or sense of self. |
First, a strong AI is an ideal AI that has emotions, self-awareness, and the ability to learn like a human being; it appears in many science fiction novels and near-future movies and is the closest thing to a humanoid “robot” in image.
For example, robots that are almost indistinguishable from humans, such as those in the movies “Eye, Robot” and the “Blade Runner” series, fall under the category of “strong AI. In the real world, however, humanoid robots are still a long way off.
Weak AIs, on the other hand, work well for specific tasks and problem solving. It is important to note, however, that it has no capabilities outside of its configured functions.
For example, the search engine Google instantly finds relevant information from countless web pages around the world, but it cannot break a single egg. Also, voice assistants such as Siri and Amazon Alexa can answer questions and perform simple tasks, but they do not think or feel like humans.
Thus, “weak and strong AI” is based on whether AI can think, learn, and understand like humans, or in other words, “how close it is to humans.
Glossary of AI Terms
The world of artificial intelligence (AI) is filled with numerous technical terms. The following table summarizes some of the most important terms.
| AI term | Explanation |
| Machine learning | How AI learns from past data to make future predictions and decisions |
| Deep learning | Computational models that are more advanced machine learning and closer to the human brain |
| Neural network | Imitation of the Brain for deep learning. |
| Natural Language Processing (NLP) | Technology for computers to understand human language |
| Reinforcement Learning | A method in which AI learns optimal behavior based on rewards. |
| Interactive and Generative AI | AI capable of interacting with humans and generating new content |
| Voice recognition | Technology for converting speech data to text and commands |
Among them, it is recommended to keep in mind the details of “machine learning” and “deep learning”, which are highly relevant to artificial intelligence (AI). Let’s take a closer look at each.
Machine learning
Machine learning is a technology in which AI automatically “learns” based on past data and uses the accumulated data to make future decisions and predictions. Its applications are expanding, especially in the business domain, and range from product recommendations to inventory management.
The key here is the “features” set by humans. A feature quantity refers to a variable used to select points or elements that require special attention from a large amount of data. For example, for a system that introduces recommended products, data such as the number of times a user has purchased a product or the number of accesses to a product page are used as feature values.
The choice of features can greatly affect the accuracy of machine learning. In other words, by selecting features well, more accurate predictions and judgments of the future can be made. The process of setting these features is called “feature amount engineering” and whether or not it is done well will greatly affect the accuracy of the prediction.
Deep learning
Deep learning is a more advanced technology of machine learning. It uses a “neural network” similar to the human brain to perform calculations.
Deep learning has higher accuracy and efficiency than machine learning because of its ability to automatically extract features. It is also the spark for the current AI boom. For example, many advanced AI systems, such as AI-based tourist attraction recommendation services and AI-powered matching applications, are operated by deep learning.
What Artificial Intelligence (AI) can do
There are many different things that can be done with artificial intelligence, but the following six are representative.
- Object recognition
- Image recognition
- Voice recognition
- Chatbot
- Diagnosis of symptoms and diseases in medicine
- Marketing Strategy Optimization
We will introduce them one by one, along with examples of their use.
Object recognition
Object recognition is a technology that uses artificial intelligence to recognize objects in the surrounding environment.
The most obvious example of object recognition is iRobot’s Roomba robot vacuum cleaner, which is equipped with artificial intelligence that uses data received from optical sensors to determine the location and layout of furniture.
Roomba, which efficiently cleans without coming in contact with furniture, has reduced human labor and revolutionized the way people clean.
Image recognition
Image recognition is a technology that uses artificial intelligence to detect and recognize specific patterns from image data.
A well-known example of the use of image recognition is the “Face Recognition Gate” provided by Panasonic. Face recognition gates have been introduced at airports around the world. Face recognition is performed by matching the face image captured by a camera with the face image data recorded in the IC passport.
Facial recognition gates have not only greatly reduced the burden on immigration inspectors, but have also facilitated smoother immigration clearance.
Voice recognition
Speech recognition is a technology that uses artificial intelligence to recognize voice characteristics and the content of words from voice data.
Siri, the well-known “Hey Siri” in Apple products, also uses voice recognition technology. Siri uses the waveform of voice data to identify the words uttered, and then converts them into text data to determine the content of commands.
Siri reduces the number of detailed smartphone operations using the hands, and many iPhone users find it convenient.
Chatbot
Chatbot is a communication technology in which artificial intelligence responds to human questions and requests.
Deep learning is used in chatbots, and the artificial intelligence accumulates natural response patterns to a wide variety of questions from users.
Diagnosis of symptoms and diseases in medicine
The use of AI in the medical field is rapidly advancing, and one example is the COVID-19 Pneumonia Imaging Diagnosis Support Program offered by Fujifilm Corporation.
The program is designed to assist radiologists in diagnostic imaging.
Specifically, it analyzes X-ray and CT images and can detect signs of COVID-19 pneumonia. This helps medical staff to quickly and accurately make a diagnosis when time and resources are limited.
Medical AI plays a role in improving the accuracy and efficiency of diagnosis, working in collaboration with physicians and healthcare professionals. This allows them to treat more patients, so patients can receive appropriate treatment faster.
Marketing Strategy Optimization
AI marketing is another innovative technology that uses data analysis and machine learning to optimize marketing strategies.
AI marketing improves the effectiveness of customer education and promotes effective customer engagement (the relationship between a company and its customers).
In addition to this, marketing automation (MA), which is now widespread in the marketing field, is another important tool to automate marketing activities and improve sales promotion efficiency. In particular, MA scenario design plans the timing and methods of sending optimized direct mail (DM) to specific targets, resulting in improved ROI (return on investment).
The six AI technologies introduced above are representative examples of items that form a new trend in marketing and are essential for companies to maintain their competitive edge.
Advantages and Disadvantages of using Artificial Intelligence (AI)
Artificial intelligence (AI) is highly anticipated in a variety of fields ranging from medicine to self-driving cars to marketing. On the other hand, there are some challenges and risks involved in the spread of artificial intelligence.
Here, we will discuss the advantages and disadvantages of utilizing artificial intelligence (AI).
- 3 Advantages
- 6 Disadvantages
Advantages of using Artificial Intelligence (AI)
There are three main benefits of utilizing AI
- Increased efficiency and reduced costs
- Advanced data analysis
- Automation of operations and 24/265 operations
AI can greatly improve work efficiency by automating tasks that do not require human intervention. For example, in Amazon’s warehouses, AI-powered robots perform picking and sorting tasks, significantly reducing manual labor.
AI can also analyze large amounts of data at high speed, which has the potential to create new businesses that utilize big data. Furthermore, AI can automate tasks, enabling 24/7 operations without the need for users.
The introduction of AI chatbots in online customer service is one example: as AI is widely applied in operations, it will accelerate on-site efficiency as human resources can be reallocated to other important tasks.
Disadvantages of Using Artificial Intelligence (AI)
The following table lists six disadvantages of utilizing AI.
The following table lists six disadvantages of using AI.
| Disadvantage | concrete example | Impact/Concerns |
| Initial cost High | System development and physician training costs are required to implement medical diagnostic AI | Large initial investment required |
| AI Data Dependency Dependency on AI data | Facial recognition systems trained on biased data cannot identify specific racial groups. | Risk of false results |
| Results Reliability | Reliability of diagnostic and other results produced by AI has not been established. | User trust issues |
| Loss of employment | Possible loss of jobs for taxi/truck drivers and others due to the widespread use of self-driving cars and drones | Risk of job loss |
| Security Issue | Example of Equifax hack: personal information of about 140 million people compromised | Risk of leakage of confidential data |
| Risk Management Increase | Potential for losses due to malfunction of AI-automated stock trading system | Increased administrative costs and effort, risk of damages |
In particular, the high initial cost is a major disadvantage.
Introducing AI requires specialized knowledge and skills and a high initial investment. For example, the initial cost of introducing medical diagnosis AI tends to be high, not only because of the system development costs, but also because of the cost of training physicians.
In addition to this, we must also be aware of security issues such as information leaks.
When AI systems are hacked, they carry the risk of leaking large amounts of sensitive data. One known example is the March 2017 cyber attack on Equifax, a major U.S. credit bureau that uses AI, which compromised the personal information of approximately ¹140 million people.
These are the advantages and disadvantages of using AI and the risks to be aware of; it is important to make a comprehensive evaluation of convenience and risk management when considering AI implementation.
How will the world change with AI?
As mentioned above, artificial intelligence is a very promising field and is expected to continue to develop. On the other hand, it will also have no small impact on human work.
The following is an explanation of the impact that the development of artificial intelligence will have on the world.
AI steals your job
49% of the working population may lose their jobs to artificial intelligence (AI)…
As artificial intelligence becomes more widespread and the world becomes more convenient, some jobs will naturally no longer need to be done by users.
According to a joint study by the Nomura Research Institute, a major think tank, and the University of Oxford in the UK, it is predicted that artificial intelligence will be able to replace about 49% of jobs in the near future.
This means that about half of businesspeople may lose their jobs to artificial intelligence.
Jobs safe from AI
First, let’s take a quick look at some examples of jobs that are likely to be lost to artificial intelligence.
First, let’s take a quick look at some examples of jobs that are likely to be lost to artificial intelligence.
- General and accounting clerks,
- receptionists,
- dry cleaning agents,
- construction workers,
- car assemblers,
- car painters,
- supermarket clerks,
- cab drivers,
- parcel delivery service delivery workers,
- train drivers,
- route bus drivers,
- customs brokers
On the other hand, the following jobs are likely to be safe from artificial intelligence
- Art directors,
- interior coordinators,
- floral designers,
- makeup artists,
- film directors,
- classical musicians,
- game creators,
- TV personalities,
- acupuncturists,
- moxibustion therapists,
- beauticians,
- childcare workers,
- manga artists,
- musicians,
- management consultants,
- industrial counselors,
- small and medium corporate diagnostician
Two characteristics of jobs that are difficult to lose to artificial intelligence are
- Requires artistic and creative skills
- Connection with people is essential
Conversely, jobs that “follow the rules to the letter” and “do not place much importance on relationships with people” are likely to be lost to artificial intelligence.
Programmer job is recommended.
We recommend “programmer” as a job that is unlikely to be lost to artificial intelligence.
This is because programmers are the only profession that can create artificial intelligence. Once you become a programmer who can create artificial intelligence, you will not lose your job but will have more jobs even if artificial intelligence develops.
Programmers are also the ones who will develop artificial intelligence in the first place, making this the perfect job for those who want to lead the way in the future.