Professor Kim Eui-seok from the Graduate School of Technology Management at KAIST has raised concerns about the limitations of artificial intelligence (AI) in research and development. As AI agents increasingly automate tasks by running thousands of virtual simulations, they are enhancing efficiencies dramatically—reportedly increasing data processing speeds by 100 times and reducing costs to one-tenth. However, these gains have not led to the groundbreaking innovations many anticipated, leaving chief technology officers and researchers feeling disheartened.
“The analysis was finished in a day, and the completeness of the results increased, so why isn’t there a ‘one shot’ that will surprise the world?” Kim stated, emphasizing that while AI excels at improving existing products through gradual innovation, it often does so within the boundaries of established knowledge.
The core issue lies in the nature of AI itself; it operates as a “probability” machine, generating outcomes based on the patterns it learns from vast datasets. While this is effective for optimization, it tends to guide companies towards the “peak of the mean” in a normal distribution curve. However, major breakthroughs that have historically transformed industries often arise from unexplained outliers, suggesting that companies may need to rethink how they utilize AI in their R&D processes.
To foster true innovation, Kim proposes a multi-faceted approach. Firstly, he advocates for a ‘noise conservation area’ where researchers actively re-evaluate data that AI might classify as irrelevant. This could potentially unearth next-generation materials or technologies hidden among what AI deems “trash.”
Secondly, Kim emphasizes the need to separate ‘exploration’ from ‘utilization’. While AI is adept at leveraging existing knowledge to improve products, human researchers should focus on formulating bold hypotheses that challenge AI’s predictive abilities. Performance indicators should then measure not just success rates, but also how far teams venture away from AI-generated predictions.
Finally, there is a call for a shift in leadership within R&D departments. Previously, R&D leaders aimed to increase the chances of success through calculated risk. Now, Kim argues, leaders must cultivate a mindset that embraces uncertainty and even embraces failure in pursuit of transformative ideas. “Drastic thinking transitions and sometimes rough intuition are needed in the context of problems that can allocate resources on dangerous paths with low probability of success,” he explained.
Kim’s insights underline a critical paradigm shift: in evolutionary terms, the most adaptable groups for survival are not necessarily the most refined but are those that incorporate variants capable of responding to changing conditions. While AI can provide efficiency and precision, it lacks the heterogeneity required to foster innovative thought.
He warns that if R&D centers are operating too smoothly and efficiently, it may indicate a stagnation crisis. “Right now, you have to throw a grain of sand of ‘inefficient other things’ into that smooth situation,” Kim advised. Innovation, he argues, often occurs where efficiency ends, suggesting that hidden opportunities, potentially worth 100 trillion won, lie within the noise of inefficient processes.
The implications of Kim’s assertions are significant for the future of technology development, as companies grapple with the balance between AI’s powerful capabilities and the unpredictable nature of human creativity. As industries evolve, they may need to embrace a dual-track approach that leverages AI for optimization while simultaneously encouraging human researchers to venture into the unknown.
As the dialogue around AI’s role in innovation expands, the need for a strategic shift in R&D practices could become vital for companies aiming to remain competitive in a rapidly evolving technological landscape.
See also
IBM Targets AI Growth in Southeast Asia While Advancing Quantum Computing Roadmap
OpenAI Seeks Head of Preparedness for $555K Role Amid Safety Concerns and Rapid Growth
AI Startups Face Rising Borrowing Costs as Debt Investors Cautiously Monitor Market Risks
Buffett’s Berkshire Bets on AI: Amazon and Alphabet Positioned for 2026 Surge
AI and Global Expansion Propel Auto Parts Sector Growth, Targeting $10B Liquid Cooling Market by 2030
