[INTERVIEW] AI advances biotech: hope or hype?

AI is expanding possibilities in biotech research, but experts warn against getting carried away with grand promises. Despite caution, there is optimism about how AI will advance life science.

Younghoon David Kim, left, chairman of Daesung Group, and Bernhard Palsson, bioengineering professor at the University of California, San Diego, speak during an interview with the Korea JoongAng Daily on Sept. 27. [PARK SANG-MOON]

Will AI pave the way for humanity to unlock the ultimate secret of life?

The question, grand as it is, has become more relevant than ever as excitement over the transformative potential of generative AI surges with equal measures of apprehension.

Over the past few decades, biologists have significantly expanded our understanding of the microscale world with new breakthroughs mapping the genome of living organisms.

Enter AI — and now the fast-evolving technology is opening new doors in the realm of bioengineering research.

However, despite the global hype, we should keep our expectations at bay, cautions the chairman of Korea’s energy conglomerate Daesung Group.

“There is a certain bandwagon effect [happening],” Daesung Group chief Younghoon David Kim, who previously served as the chair of the World Energy Council, said. Though not pessimistic, the chairman underscored prudence and practicality in approaching its grand promises.

Nonetheless, there is also solid optimism about how AI could lead to progress in the field of life science.

According to Bernhard Palsson, professor of bioengineering at the University of California, San Diego, “by the end of this decade, we will probably be building genomes that are maybe a million, or a million and a half base pairs,” which is equivalent to that of small bacterial genomes.

Palsson, a pioneering expert in genome-scale models, highlighted how AI technology is catalyzing major shifts in data analytics within life science, where the exponential growth in data types demands advanced analytic tools.

Daesung Group, one of the oldest energy conglomerates in Korea, has been hosting the annual Daesung Haegang Microbes Forum since 2017 with leading experts in science.

This year’s edition, held on Sept. 26, was themed “The Impact of AI on Biotech,” with Palsson attending as a plenary speaker.

On the day after the forum, the Korea JoongAng Daily sat down with Kim and Palsson at Daesung Group’s headquarters in central Seoul to discuss the hope and the hype surrounding AI technology and its potential for real-life applications in bioengineering.

This interview has been edited for length and clarity.

Q. Biotechnology and AI are gaining significance in Korea's national agenda, as well as on a global scale. What, in your opinion, is the major driver behind the Korean government's focus on the bioeconomy and AI technology?

A. Kim: I think there is a certain bandwagon effect happening. Every government is trying to catch up with this trend in AI and biotechnology, yet there was a huge fever for biotechnology about just 10 or 20 years ago, and that has cooled down.

I think the same goes for AI as well. When we first discovered the genome world, people were fascinated by its possibility — but the progress made over the past 50 or 60 years wasn't that big compared to what we expected.

So even though I'm looking into the possibilities, the promises seem a little exaggerated. As a businessman, I'm a bit cautious about the progress of both biology and AI.

Younghoon David Kim, chairman of Daesung Group, speaks during an interview with the Korea JoongAng Daily on Sept. 27. [PARK SANG-MOON]

AI technology has witnessed significant advancements in recent years. How has the recent progress in AI models influenced the field of bioengineering?

Palsson:DNA sequencing became very cheap around 2010. For about 10 years after that, people sequenced a lot of genomes and transcriptomes. So we had massive growth in available data, and that demanded analysis.

And it turned out that many of the algorithms that have been developed in the AI field could be used to analyze that data, so we made a lot of progress with analyzing individual data types over the last five to seven years using existing algorithms.

As the global AI race picks up pace, how would you assess the progress made especially in the context of bioengineering applications?

Palsson: I've been doing biotech for 40 years now, and there are always these waves of hype and disappointment. The hype starts with a new algorithm or a new experimental method, and when you look at the science it often leads to a better understanding or explanation of a phenomenon. But when it comes to the practical utility, it gets much tougher, And that's where a lot of the failures are.

With AI and the information technology revolution, it is very clear that we have a lot of information available to us now than before. If you properly analyze the information, it leads to understanding, or maybe knowledge, which enables prediction and design.

One specific example that is gaining a lot of traction is to sequence human pathogens and use them to track how a disease spreads and develops. As we saw during the Covid-19 pandemic, DNA sequencing could allow us to track how this virus changed and how it moved around.

I would say the same thing is about to happen for bacterial pathogens. I think over the next 10 years we will build screening capabilities in hospitals around the world to monitor what pathogen is where, and how it moves around, mutates, adapts. This is observation, which will be helpful for physicians in deciding how to treat a new pathogen found in a patient.

Bernhard Palsson, bioengineering professor at the University of California, San Diego, speaks during an interview with the Korea JoongAng Daily on Sept. 27. [PARK SANG-MOON]

The rise of generative AI has ignited many new discourses surrounding the ethical use of AI and its environmental impact.What are some major challenges that AI is posing in this era and what are some possible solutions for that?

Palsson:To pick one issue, as you mentioned, the technology is going to use a lot of energy and we hear this always about Bitcoin mining. There is a whole field now on what they call low-power computations, which started with satellites that go into space with little power available. So people have been trying to design chips for low-power computation, and as we have seen in the previous day's forum, chip architecture can mitigate such power problems.

Kim: Decades after Richard Feynman proposed the idea of quantum computing, we still do not have a commercially viable option. I am not saying that all the promises and enthusiasm are necessarily destined for failure, but many scientists and companies will face quite a bit of them. And so I myself am divided for most of the possibilities.

These advanced research projects may belong to the public sector, rather than the private, but the problem is that those receiving the government subsidies can be less responsible, and now the government is reviewing the process.

I agree with the sentiment that there is a new world coming with AI, but human history is full of contradictions. AI is based on quadratic equations, so we basically rely largely on binomial quadratic equations when it comes to the technology. But everyday life in business tells me that life is not quadratic at all — it is very erratic.

How did the annual Daesung Haegang Microbes Forum start, and what is the significance of the conference?

Kim: We have a landfill gas station, where we are taking advantage of anaerobic microbes that fly in the air before they settle down in our site and start producing gases. But as years went by, its productivity declined. So I try to find, or hunt for, new microbes that can further enhance productivity.

I ended up talking to microbes experts during the World Economic Forum, who introduced huge possibilities even beyond gas production. If I can mobilize these small microbes effectively, I will be able to do many other things, or so I hope, and that's why we began to hold the Daesung Haegang Microbes Forum.

What was the most significant takeaway at this year’s forum?

Palsson: AI has been around for a long time — and I was part of an AI center in 1990 funded by NASA. But the technology has been developing for a long time, and we all know that now Google solved what's called the next-word-prediction problem. So if you’re building a sentence, you can keep predicting the words, a capability that shocked everyone.

And we're all thinking about how we can take those natural language models and use them for biological objects rather than language. For instance, with Google’s AlphaFold 2, we can now predict protein structures based on data that we have.

The question that needs to be answered now is how broadly applicable will AI be for many other biological functions, and there's a lot of optimism, a lot of hope. I think it will take a little longer than many expect, but there is definitely a big change happening right now in data analytics driven by AI.

Daesung Group has been investing in content production and emerging industries such as quantum computing through Daesung Private Equity. What is your major focus in making investment decisions?

Kim: I cannot explain everything, but one thing that is important is whether it can be repaid in a promised time span. In terms of the content industry, in particular, we try to focus on family-friendly content, which can be viewed with all family members gathered together, as there is a lot of violent, sensational or harmful content in the industry. And we have been very successful with content investments, some examples being “The Admiral: Roaring Currents” (2014), and “Ode to My Father” (2014).

For investments in high-tech industries, which are very close to our business needs like synthetic biology, I don't want to be drifted by all the hype, so I take a step-by-step approach. Some people in this industry can be a little irresponsible, I think, as they'd put in a lot of money based on speculations. We stay away from such tendencies, though our performance may fall behind compared to such investors. But in the long run, I think we are right.

BY SHIN HA-NEE [shin.hanee@joongang.co.kr]