Addressing the corporate AI dilemma

Lee Kyung-bae
The author is an adjunct professor at the Institute of Convergence Science,Yonsei University.

When generative artificial intelligence (AI) programs such as GPT-3.5, GPT-4.0 and SORA — a video-generating AI model — were released about a year ago, their innovative capabilities were very impressive. Now, recently launched technologies such as OpenAI’s GPT-4o, Google’s Gemini-1.5 Pro and Veo, a video-generating AI model, are astonishing. The competitive releases of such robots and smartphones with cutting-edge AI technologies are not only sparking users’ curiosity, but also heralding revolutions between countries and across industries.

Let’s take a look at a recent survey of more than 30,000 people in 31 countries on the use of generative AI. Three out of four employees, or 75 percent, say they are using AI technologies for work, and 85 percent said they personally use AI tools without company support. While 79 percent of employees in leadership positions said AI is essential for their companies to maintain competitiveness, 60 percent said they worry about a lack of vision and clear plans about using AI for their work.

Companies are moving fast to apply generative AI to their work, but are also utterly perplexed as novel AI models equipped with dazzling features are released every day.

While individual users can easily use AI technologies with a little training, companies must take into account many things when they apply the technologies to company operations. There are several steps companies can take to use AI for their business.

The first step is the “personal assistant stage,” where AI is applied to groupware, document creation and other works. The second step is the “search assistant stage,” where AI queries and analyzes internal and external laws, regulations, standards, guidelines, tax codes, manuals and the current status of business performance.

Third is the “business assistant stage,” where AI processes and manages tasks in conjunction with the company’s legacy systems. The fourth is using AI as a “management assistant” that connects data and systems at home and abroad to manage and support decision-making. There is also the fifth stage where AI makes its own judgment and executes plans, but it will likely be safer if the technology is still under human control than otherwise.

What should we do to introduce such AI systems for companies?

First, it is necessary for a company to build consensus on how to apply AI to their work through workshops, for example.

Second, the company should develop an information strategy plan (ISP) based on the outcomes of the workshops. You can expect great results from an AI service when high quality data are systematically organized. Therefore, the ISP must include a set of plans to secure data from inside and outside the company, clean them and upgrade information technology systems with the data.

As companies have just started using AI, they do not have solutions or the manpower to deal with system connections, data management and information security. Therefore, it’s better for companies to apply AI to their operations step by step rather than changing the entire system from the start.

Third, it is necessary for companies to verify the effectiveness of applying AI through the proof of concept. Based on this, companies can select the most suitable AI product and decide whether to build data on their own or use the cloud.

Fourth, companies must designate an AI specialist in each department to push forward digital innovation using AI.

Fifth, as many new technologies are being developed and released, it is necessary to strengthen an executive-level innovation team capable of pushing forward a project to integrate information technology, digital transformation, big data, AI and digital-related tasks, which are scattered across various divisions of the company. This innovation team should also be awarded with governance power.

It’s not wise to use such powerful generative AI simply to replace manual labor or take meeting minutes. Companies should make bold decisions to use AI to do things that would otherwise be difficult or apply the latest management practices and production theories. Companies can also act boldly to use AI to analyze vast amounts of data to increase productivity or develop new products that reflect trends.

Introducing AI to improve individual workers’ productivity should be a secondary goal.

Translation by the Korea JoongAng Daily staff.