Main menu


What is the future of artificial intelligence?


What is the future of artificial intelligence?

Future of artificial intelligence AI

Technology is like a constantly moving living organism. Just as a person starts from the beginning, they grow, sometimes they are pushed out of the competition and eliminated. Let's look at the changes related to AI in the future.

Technology is like a constantly moving living organism. Just as humans start at their infancy, technology also emerges in a poor state and grows to a high level through innovation, or it is pushed out of competition with other technologies. The way the market embraces technology is also constantly changing. As soon as it appears, the public pours a lot of interest, and although the initial market is formed, it often takes a long time to spread. In order to devise a strategy to create a big impact through AI, it is necessary to understand the mechanism and long-term direction of how technology and market environment change. In that sense, let's take a look at the changes related to AI in the future.

Future of artificial intelligence AI

1. The exponential: Accelerating technological advancement Artificial intelligence AI

Above all, AI technology will advance at an even faster pace. The MIT Technology Review has consistently introduced AI technologies that are constantly evolving. Moore's Law is a concept commonly used when talking about the development of IT technology. This means that the number of transistors in a semiconductor circuit doubles every 1.5 years. The rate at which the technological performance of AI is advancing is projected to be 5 to 100 times faster than Moore's Law (ARK Investment, 2020). The computational processing power of AI learning models is growing tenfold every year, so this exponential growth will continue in the future.

2. Artificial intelligence AI Data Big Bang: Explosive Data Growth

Second, the amount of data is growing rapidly. Data is an essential material for AI modeling. Worldwide, data is growing exponentially. Data Never Sleeps, which analyzes data statistics, presented the amount of data produced per minute through various applications and services in April. YouTube is pouring out a lot of data with 500 hours of video uploaded every minute. About 147,000 photos are uploaded to Facebook and 150,000 messages are shared, and about 41 million messages are shared per minute on WhatsApp, an SNS service. In the case of Zoom, more than 200,000 people connect for meetings per minute, TikTok has 2,704 people installing the application, and Amazon is shipping 6,659 products per minute. A significant amount of data is produced among these activities, and the amount is increasing every year.

3. Reduction of cost: reduction of Artificial intelligence AI training cost

The third is cost. Advanced AI models such as deep learning provide excellent performance, but the disadvantage is that they are expensive. For deep learning training, high-end hardware must be prepared, and it takes a long time to learn massive data. The amount of manpower that is put in to prepare data and carry out modeling is not formidable. However, as hardware technologies such as GPU and TPU are developing recently and more efficient data processing methods have emerged, the learning cost is decreasing to 1/10 a year.

4. Heterogeneous bonding: bonding between technologies in other fields artificial intelligence AI

The fourth is that the combination of technologies will become more widespread. AI can create added value by being compatible with various technologies. For example, Metaverse, which has recently become an important topic in the industry, can be combined with AI. Metaverse is a compound word of meta meaning virtual and transcendence and universe meaning real world, and refers to mixed reality where virtual and real interact. AI can create an avatar in the metaverse similar to the user's real appearance, or it can be designed to communicate with the user by applying a natural language processing-based communication model. In addition, it can recognize users in virtual reality and show customized advertisements. AI will play an important role in making this metaverse world intelligent. In addition, it can be combined with blockchain to establish an intelligent security system or to predict the volatility of household currency to improve asset stability. The 4th industrial revolution is a time when advanced new technologies appear simultaneously. In this situation, AI is expected to play a role in promoting disruptive innovation through combination with other technologies.


5. Proliferation: Expanding the scope of  artificial intelligence AI adoption

Fifth is the expansion of the scope of application of AI. Due to the great potential of AI, many industries are adopting this technology. Currently, AI is being actively introduced in fields such as life science, distribution, and consumer goods. So far, it is generally at the pilot or proof-of-concept (PoC) level. However, in the future, it will be applied to all industries and is expected to mature in the direction of being applied to actual products and services.

6. artificial intelligence AI Universalization: universal use of technology

The sixth is the generalization of AI technology. Although AI technology has a long academic history, it has only recently entered the industry. As a result, it is true that this technology is unfamiliar to many companies now. However, it will not remain as uncharted territory in the future as it is now. In the early 1990s, when the Internet boom occurred, many companies introduced web services and e-commerce. This is because it was expected that new added values ​​would be created in this area. However, in a situation where everyone is using the Internet, the Internet itself cannot make a difference now. The same goes for AI. As the potential is great, many companies will adopt this technology, and over time it will become a general-purpose technology. In other words, introducing AI itself does not guarantee differentiation. Currently, there is a shortage of AI technical talents, but the AI ​​graduate school and various educational programs are set up to produce many talents, so the talent shortage problem is expected to be gradually resolved. Of course, talent with both technical and industry domain knowledge will continue to be scarce.


7. Impact Gap: The Competitiveness Gap Deepens artificial intelligence AI

The seventh is that the gap in impact created by AI is widening among companies. The introduction of AI does not mean that the same impact can be created. Few of the real AI adopters know exactly how AI works and are linking it to distinct value creation. According to a McKinsey study, mature companies that successfully introduced AI achieved 3 to 4 times higher operating profit than latecomers. These companies account for about 10% of the total. On the other hand, about 30% of immature companies that have introduced AI but have not yet used it effectively, although they create an impact, only produce an operating profit effect of 1.8 times that of latecomers. In other words, although the introduction of AI can have an impact, there can be an impact gap depending on how effectively AI is used. Over time, this impact gap is expected to widen.

Summarizing these changes, the technological performance and cost-effectiveness of AI will be further improved, and it will be applied to a wider range of industries through technology convergence. However, as the technology becomes commoditized, the benefits of AI in itself will gradually disappear. In the future, the difference in competitiveness will widen depending on how effectively AI has been introduced and created a lot of impact. From now on, it should be seen that the task to be focused on has been decided. It is to have a strategy to flexibly absorb the increasingly advanced AI technology into its business and create a distinct impact through it. In order to have such a strategy, it is necessary to understand the types of impact that can be created through AI, along with an understanding of the technology itself, and how to lead from the actual AI introduction to the impact.

Jung Doo-hee is the editor-in-chief of MIT Technology Review Korea, and is a professor in the Department of ICT and Entrepreneurship at Handong Global University. She serves as a representative partner of Impactive AI, an AI consulting company, helping Korean companies successfully introduce AI. She is the author of <AI Business Model Ended in One Book>, <The AI ​​Hypergap Era is Coming in 3 Years>, and <TQ Technology Intelligence>.