'AI will have significant impact across all industries': McKinsey

Matt McDevitt, McKinsey partner and co-leader of QuantumBlack Japan, speaks during the Korea Times Forum on “Survival and Growth Strategies in the AI Era” at the Korea Chamber of Commerce and Industry (KCCI) building in central Seoul, Wednesday. Korea Times photo by Shim Hyun-chul

OpenAI’s ChatGPT allowed many people to experience the revolutionary changes that generative AI can bring to our lives. This emerging technology is forecast to continue to develop and have a major impact across all industries, bringing about many changes in our lives, a partner at global consulting company McKinsey & Company said during the Korea Times Forum titled “Survival and Growth Strategies in the AI Era,” Wednesday.

The McKinsey partner also predicted that generative AI will unlock $4.4 trillion in annual value for the global economy, and that the technology will boost businesses across a wide range of sectors, from advanced technology to manufacturing, energy, education and retail.

“We recently released a report on the economic productivity potential of generative AI, and it is estimated that there’s $4.4 trillion in economic potential. This is not even tapping into the potential recommendation, things that we cannot predict and things that we have not seen thus far. This is only around looking at industries around the world,” Matt McDevitt, McKinsey partner and co-leader of QuantumBlack Japan, said during his keynote speech at the forum.

The English daily held the forum together with McKinsey & Company to delve into the theme of “Survival and Growth Strategies in the AI Era” to focus on why this technology is becoming more important, how we utilize this technology in various sectors and what we need to do to improve the country’s AI capability.

 

At the forum, prominent figures in the AI sector and policymakers had discussions about the impact of generative AI on the industry and how this rapidly developing technology can be used positively without being abused.

During his speech, the McKinsey partner said that there have been significant changes in the perception of generative AI in the past few months since OpenAI released the 3.5 version of ChatGPT last November.

Before the advent of the 3.5 version, there were many questions about what generative AI is and whether it is realistic. However, now there are discussions about how to apply and use this technology in various industries, such as where and how should we start with generative AI.

“Just a few months ago, people are asking, what is gen AI? Is it all hype? Or is it all reality? As you just saw this earlier, it can be used to write speeches for people and it basically looks and feels like a human being,” he said. “Today we're asking a lot of different questions. We're asking not just what generative AI is, but where we should apply it, how we should leverage it, how we ensure that risks are mitigated and how we partner with organizations to be able to move quickly and efficiently in this space.”

He also introduced the efforts of global companies to develop and use generative AI in their own ways. “Morgan Stanley, for example, has announced that they're using AI to support their wealth managers globally around doing their work,” he said.

In the life science sector, Insilico Medicine developed a generative AI model to predict clinical trial success rates with over 80 percent accuracy, and AstraZeneca accelerates drug discovery by training AI models on the grammar of biochemistry and digital pathology images to help generate new molecules.

In the computer software sector, Adobe has developed a suite of generative AI tools to enhance its image and video-editing software platforms with generative capabilities, and Naver also launched HyperCLOVA X using a Korean LLM and also launched Cue:, its generative AI-based search engine. Another Korean company SK Telecom launched A., its own Korean large language model and generative AI based personal assistant service.

“We've seen examples of this in Korea. HyperCLOVA X is able to focus on the Korean language make a Korean large language model more efficient,” McDevitt said.

Matt McDevitt, McKinsey partner and co-leader of QuantumBlack Japan, speaks during the Korea Times Forum at the Korea Chamber of Commerce and Industry (KCCI) building in central Seoul, Wednesday. Korea Times photo by Shim Hyun-chul

Organizations need to understand their position to use AI effectively

He defined the value of generative AI in terms of the four Cs ― concision, which is to summarize and extract insights; coding and software, which interprets and generates code; creative content, which generates texts, images and others; and customer engagement, which provides customer service through chat and expands chatbot usage toward client outreach and data collection.

Giving advice to organizations that want to utilize and invest in generative AI, McDevitt added they should start by identifying whether they are a taker, shaper or maker.

According to him, a taker is a company that simply applies existing generative AI to its workflow without customization. A shaper is a company that uses its proprietary data and insights to enhance existing generative AI to fit its organization and business needs. Alternatively, a company can be a maker that develops a completely new model tailored to its specific needs.

“When we think about organizations and their decision to strategically invest in generative AI, there's effectively three areas where people can take a stance. There is the taker situation in which you are taking off the shelf technology applications waiting for the technology and industry to mature. There is shaper which is when you actually modify the technology. And then there’s maker. These are companies that are like Google and Amazon and Uber that are investing in large language models and generative AI technology,” he said.

“We find that almost all the organizations that we interact with are not going to be makers. All of them will at least be takers, but almost all of them also are considering how to be shapers and the investments required to the effective and efficient around this area is really a key.”

He added that the key thing organizations should consider is not the kinds of generative AI services, but data architecture to improve fundamental competitiveness.

“Organizations have consistently, continuously needed to invest in data architecture to improve their underlying technology,” McDevitt said. “Data is not a sexy thing to invest. It is kind of like cleaning up your room and doing chores, but it's a critical element that they don't do correctly.”