Artificial intelligence (AI) had previously considered creative activities such as writing, drawing, and coding to be activities that only humans can do. However, with the development of artificial intelligence (AI), it is now the time for Generative AI to look beyond the human realm. Now, Generative AI is leading the transition to a new technological paradigm, starting with ChatGPT.


What is Generative AI?

Generative AI, which means generative artificial intelligence, generates original and realistic content such as images or texts without humans entering large amounts of data. OpenAI's ChatGPT played a big role in the reason Generative AI became a hot topic. ChatGPT has been used by 100 million people a month in two months since its launch last year, and is now used by 100 million people a week.


Change brought about by Generative AI?

Work productivity improvement
More and more businesses use ChatGPT to increase productivity. Some 92 percent of Fortune 500 companies use ChatGPT. Microsoft also recently released Generative AI Copilot in the form of ChatGPT. Not only does Copilot write code on behalf of MS programs, but it also works on Word, Excel, and PowerPoint in conjunction with MS programs. Some predict that the gap in work productivity between workers who use these generated AI technologies and those who do not will widen in the future.

A chatbot who doesn't know how to code
Anyone, not even a developer, can now create AI chatbots. OpenAI released a service called GPT Builder on Nov. 9, which allows users to create customized AI chatbots. By setting the desired purpose and entering materials for AI to learn about the topic, they can create their own custom chatbots without coding.


GPT-4 Turbo

On Nov. 6, OpenAI announced its latest AI model, GPT-4 Turbo, at a conference. The GPT series is a large-scale language model on which ChatGPT is based, and the GPT-4 Turbo offers significant performance improvements over the previous model. Key improvements to the GPT-4 Turbo are as follows.

1. Up to 16x the amount of information you can enter at a time
2. Can analyze the content of uploaded images and generate image responses
3. Allows text to be made natural audio
4. Fine-tuning and introducing a copyright protection system for businesses

In addition, it is trained with data up to April 2023, allowing up-to-date answers over the existing model GPT-4, which was learned with data up to September 2021. In addition, OpenAI has also unveiled GPTs, a service that allows ordinary users to create their own ChatGPTs themselves. GPTs can be programmed in everyday language without coding, and OpenAI has announced that it will upload the service to the GPT Store for transaction.

Google's Generative AI-Based Search Service

Google also announced that it will expand its Generative AI based Search Generative Experience (SGE) to 120 countries, including Korea. Following Google's interactive chatbot "Bad" supported Korean from its first launch, Generative AI's search service is now available in Korean. Generative AI search is a new feature that allows users to check information directly by displaying a link from the source along with the search results.

Matter of Generative AI

Due to the popularization of Generative AI, not only companies but also the general public have more access to AI. However, Generative AI is creating various social problems such as privacy issues and criminal use.
In particular, there is a growing risk of hacking and personal information leakage using AI. Open AI leaked personal information such as names, e-mail addresses, and credit card numbers of some users around the world who accessed ChatGPT in March, and the leaked information included information on 687 Korean users. In response, the Personal Information Commission fined OpenAI for violating its reporting obligations that did not report within 24 hours of recognizing the leak, pointing out that it was insufficient to comply with its obligations under the domestic protection law.
AI phishing scams using AI are also popular. AI phishing is a method that uses AI technology such as deep voice and deepfake to falsify faces and voices into victims' acquaintances and then trick the other person into stealing money. The threshold for crime using Generative AI is lowered because it is not difficult to create human voices and videos.

In this situation where AI has penetrated into life, the U.N. has adopted a resolution titled "Human Rights Standards in the Age of AI" to consider the human rights implications of AI technology and prevent side effects caused by reckless use. Responsible evaluation of the impact on AI human rights is needed, and AI technology should have transparency and accountability.
In the EU, companies that want to build generative AI systems such as ChatGPT by passing the AI Act must disclose the source and copyright of the data they put into machine learning, and if they want to sell AI services, they must first submit them to the EU before launch to be investigated for risk. Once the legislation is signed, regulations will be applied to EU countries from 2026.



Prospects for generative AI market
IDC, a market research firm, predicts that the global AI market, including super-giant AI, will reach USD 554.3 billion (W700 trillion) by 2024. The domestic AI market is also expected to grow by an annual average KRW 3.66 trillion in 2023, reaching an annual growth of KRW 4.4636 trillion by 2027. In addition, the demand for generated AI is expected to continue to grow as AI is accelerated in various industries. The Generative AI is already having a great impact on our lives and playing an important role in greatly improving work efficiency and information accessibility. However, it seems that there are still issues to be solved, such as ethical issues, criminal abuse issues, and information security issues.