Recently, as OpenAI and Anthropic have argued for slowing AI development speeds to prepare for AI risks, discussions on AI safety concerns have intensified. In response, Deputy Prime Minister Bae Kyung-hoon, who also serves as the Minister of Science and ICT, stated his position: "What matters is not speed control, but speed responsibility."
On the 16th, Deputy Prime Minister Bae posted on his social media (SNS) regarding the heated debate within the AI industry over slowing development speeds. He said, "From South Korea's perspective, one thing is clear: We have no luxury to slow down our AI speed at this moment."
He further emphasized, "We must go beyond creating independent AI foundation models based on LLMs (Large Language Models) and expand into technical areas such as VLMs (Vision-Language Models) and VLAs (Vision-Language-Action Models), which can be observed, understood, and acted upon in industrial settings. Moreover, we must prepare for an era where AI is deeply integrated into all industries and daily life."
Following the recent launch of OpenAI's 'Astra,' evaluations suggest that the AGI (Artificial General Intelligence) era, characterized by human-level intelligence across all domains, is imminent. In this context, securing frontier AI capabilities is seen as a decisive factor in national industrial competitiveness, economic strength, security, and future global competitiveness.
Deputy Prime Minister Bae stated, "The issue of a leading country adjusting its speed is entirely different from that of a chasing nation that must simultaneously create new markets slowing down on its own." He added, "This does not mean we should rush blindly; rather, as AI accelerates, our pace in preparing for safety and trust must also increase."
This means technical and institutional safeguards to address issues such as deepfakes, cybersecurity, personal information protection, and job market shifts must advance alongside AI development. He views the Basic Act on Artificial Intelligence Development and Trust Foundation, which took effect this year, as a critical brake mechanism against potential risks.
He said, "Speed is competitiveness, and safety is sustainability; thus, choosing one over the other is not an option." He further noted, "We will ensure seamless preparations for implementing the AI Basic Act, developing 'Dokpamo' models, promoting 'AI for All,' launching security-specialized model projects, and advancing frontier AI model initiatives to propel South Korea into a true top-three AI nation."
He also added, "South Korea will not delay its AI timeline; we will move faster and more safely."
in TECH
