"Record Productivity Increase" Is the Era of AI Creating New Drugs Coming? [Im Ju-hyung's Tech Talk]

Increasing Attempts at AI-Based New Drug Development
'Blockbuster Drug' Cost Burden High... AI as a Breakthrough
Automating 'New Drug Discovery' Stage Using AI
"In 10 Years, AI Might Create All Medicines"

Attempts to develop new drugs using artificial intelligence (AI) are increasing.
Photo by Yonhap News

[Asia Economy Reporter Lim Juhyung] In February last year, the AI-based pharmaceutical company 'Exscientia' attracted attention as the world’s first to enter human clinical trials with an AI-based new drug candidate. Not only was it the first time an AI-selected drug candidate reached the clinical stage, but the time taken for AI to discover this compound was also record-breaking. It took less than 12 months.

Applying AI to drug development allows for much faster and cheaper drug creation compared to the traditional process conducted solely by human scientists. It has the potential to completely overturn the pharmaceutical industry landscape.

The UK-based emerging startup Exscientia has partnered with the Japanese pharmaceutical company 'Dainippon Sumitomo Pharma' to begin clinical trials for an obsessive-compulsive disorder treatment.

According to the company, it usually takes an average of 4.5 years to develop a new drug and reach clinical trials, but the drug developed by Exscientia reached the clinical stage in just one year. This not only significantly shortened the development period but also reduced the research and development (R&D) costs, which run into trillions of won.

Exscientia, which earned the title of 'the world’s most promising AI drug development company,' was listed on the US stock market Nasdaq last month, raising over $500 million (approximately 592 billion won), marking one of the largest biotech IPOs this year. This indicates growing investor interest in AI drug development.

AI can dramatically shorten the time and cost required for new drug development. / Photo by Yonhap News

Companies attempting AI drug development are not limited to startups. Well-known global pharmaceutical giants such as Pfizer, AstraZeneca, and Roche are also collaborating with tech companies to incorporate AI into their R&D efforts.

For major pharmaceutical companies, AI drug development is seen as a new breakthrough. This is because the cost of developing new drugs has become so high that creating new medications has become increasingly difficult.

Typically, global pharmaceutical companies earn revenue by developing new drugs and licensing them to generic drug manufacturers and others. Among these, the drugs responsible for pharmaceutical companies’ sales are so-called 'blockbuster drugs,' products that can generate annual sales exceeding 1 trillion won.

However, blockbuster drugs require enormous development costs and carry high risks of failure. There is a possibility of losing trillions of won in R&D expenses without any returns.

The difficulty in developing blockbuster drugs can be divided into two main reasons. The first is the 'drug discovery' process, which involves selecting the most promising candidate from thousands of potential new drug compounds, and the second is the complexity of clinical trials that test the discovered drug on humans.

'Blockbuster new drugs' are major products responsible for the sales of global pharmaceutical companies, but the problem is that they have a high risk of failure. / Photo by Yonhap News

According to the 'Current Status of New Drug Development and Analysis of New Drug Development Costs' report by the Biotechnology Policy Research Center, developing a blockbuster drug takes an average of 10 to 15 years and costs about $1 billion (approximately 1.184 trillion won).

AI can take charge of the drug discovery stage in this complex drug development process. Instead of scientists analyzing nearly 10,000 new drug candidates one by one, AI armed with big data filters and recommends candidates. This can dramatically shorten the drug discovery process, which accounts for about one-third of the total drug development R&D budget and takes more than 3 to 4 years.

Scientists building AI for drug development predict that AI will soon transform the pharmaceutical industry itself.

Andrew Hopkins, founder of Exscientia and professor at the University of Dundee in the UK, said in an interview with the BBC, "To develop a new drug, billions of decision-making processes must be performed, but thanks to AI, we were able to test only 350 compounds, which is one-fifth of the typical new drug compound candidates," emphasizing, "This is a record productivity increase."

He added, "This is the first time a drug designed by AI has been created. However, in 10 years, AI might be able to create all medicines."