[Edaily Reporter Yoo Jin-hee] Global capital markets are growing increasingly wary of so-called “AI washing.” As companies that have lured investors by touting AI technology find themselves in the crosshairs of judicial authorities, the domestic pharmaceutical and biotech industry is also beginning to seriously distinguish the wheat from the chaff among AI firms. Analysts suggest that pharmaceutical and biotech companies that have merely used AI as a label—relying on generic models without developing their own proprietary engines—will ultimately face the risk of being forced out of the market.

Warnings of ‘AI washing’ from overseas… Domestic firms also under scrutiny for ‘disclosure violations’

According to the AI industry on the 27th, these market shifts first began in the United States, a country that has long driven changes in information technology (IT). Local tech firms such as Delphia and Global Predictions, which were fined hundreds of millions of won by the U.S. Securities and Exchange Commission (SEC), are cited as prime examples.

These companies sounded an alarm for the market after being caught falsely promoting the use of proprietary AI-based investment strategies. This is interpreted as a result of such practices being defined not merely as marketing exaggerations, but as deceptive acts that disrupt the order of the capital market. This sense of crisis is now spreading directly to the biotech industry.

In 2023, the U.S. AI-bio company Absci announced “zero-shot” technology, which designs antibodies using only AI without any prior data. However, it became embroiled in controversy over whether it had actually relied on existing data, as questioned by industry experts. Following the zero-shot controversy, Absci has found itself in a position where it must constantly prove its worth, such as through the announcement of a new platform (Origin-1).

The situation in Korea is no different. While some pharmaceutical and biotech startups claim in public disclosures that “AI adoption has shortened new drug development time by five years,” analyses suggest that many still rely heavily on traditional experiments. In fact, among unlisted companies, there is a growing number of cases where firms fail to even clear the threshold for technology evaluation.

Consequently, experts point out that simply utilizing external application programming interfaces (APIs) like ChatGPT or Gemini without possessing core proprietary algorithms, while packaging them as proprietary technology, clearly risks violating disclosure regulations.

An AI industry insider pointed out, “Just as we don’t call a company an IT firm simply because it uses Excel or Word, classifying companies that use ChatGPT to summarize papers or assist in compound searches as AI biotech firms is misleading to investors.” They added, “Only companies that possess proprietary learning models for designing new substances or predicting protein structures—and have secured patents for them—should be allowed to use the ‘AI biotech’ label.”

Authorities are also aware of this and plan to block AI-washing companies from entering the investment market. For example, under the strengthened KOSDAQ technology-based listing guidelines effective this year, passing the preliminary listing review is virtually impossible based solely on the description that a company “provides AI services.”

Since January of this year, the Korea Exchange has significantly strengthened its review of the AI industry, specifically requiring proof of the distinctiveness of core algorithms, the quantity and quality of training data, and bias mitigation technologies. In particular, companies that merely utilize open-source models will receive very low ratings in technology evaluations. The authorities view that pushing for an IPO without sufficient technological capabilities could lead not only to a decline in corporate reputation but also to the risk of being designated as a managed stock or delisted.

Nevertheless, Korea’s AI Competitiveness Remains Strong... ‘Genuine’ Players Like Lunit and Syntheca Bio Are Making Their Mark

Nevertheless, there is a growing consensus that the AI competitiveness of Korea’s IT-driven pharmaceutical and biotech sectors remains world-class. According to a 2026 report by the global analytics firm Deep Knowledge Group (DKG), South Korea’s biohealth AI competitiveness ranks 11th globally, placing it among the top tier.

Companies classified as ‘gems’ are already proving their value through their performance. #Lunit surpassed 10 billion won in revenue last year through its AI biomarker platform, Lunit Scope, recording a growth rate of over 100% for three consecutive years. Its integration into the companion diagnostic pipelines for anti-cancer drug clinical trials through collaborations with global big pharma companies such as Daiichi Sankyo and AstraZeneca is viewed as a unique competitive advantage that Big Tech cannot replicate in the short term.

#SynthecaBio has also gained an edge in the efficiency race by building its own large-scale data center—the only one in Korea to house a cluster of over 5,000 CPUs and GPUs. Jeong Jong-seon, CEO of SynthecaBio, stated, “AI-driven drug candidate discovery ultimately comes down to a race between large-scale screening capabilities and the speed of experimental validation,” adding with confidence, “This year will be a turning point in terms of performance as well.”

#Pharos iBio is smoothly progressing clinical trials for “Rasmotinib” (PHI-101), a treatment for acute myeloid leukemia, through its Chemiverse platform, which has been recognized for its technological prowess to the extent of collaborating with NVIDIA. The company recently demonstrated its technical capabilities once again at the American Association for Cancer Research (AACR) 2026 conference by confirming synergistic effects when combined with a global MENIN inhibitor.

Experts diagnose that, ultimately, to quell market turmoil and ensure funds flow to truly innovative companies, investors’ discernment must also improve.

Hong Soon-jae, CEO of BioBook, stated, “The three key standards for AI-driven pharmaceutical and biotech ventures are: △ whether they possess a proprietary platform capable of identifying targets using their own algorithms without relying on external models △ a validation system that rapidly feeds back AI predictions into actual experiments, and △ proof of efficacy through global technology licensing (L/O) and clinical trials,” he explained. “In addition, examining whether AI models have been validated in authoritative journals such as Nature and Science, or at global AI conferences like ICLR and NeurIPS, can also serve as a method to distinguish the wheat from the chaff.”