South Korea has embarked on an ambitious regulatory journey, unveiling what it claims to be the first comprehensive artificial intelligence (AI) laws in the world. These pioneering regulations aim to set a global standard for AI governance, yet they face significant backlash from both technology startups and civil society organisations, highlighting a divide in perspectives on the appropriate level of oversight.
A Bold Legislative Move
The new legislation, known as the AI Basic Act, came into effect last Thursday, responding to mounting global concerns regarding the implications of AI technologies. As nations grapple with the rapid advancements in AI-generated content and automated decision-making, South Korea seeks to position itself as a leader in the field, aspiring to join the ranks of the United States and China as a top AI power.
Under the act, companies will be required to implement measures such as digital watermarks on AI-generated content, with specific provisions for identifying deepfake materials. High-impact AI applications, which encompass systems used in critical areas like medical diagnostics and hiring, will mandate operators to conduct thorough risk assessments and maintain detailed records of their decision-making processes. Notably, the legislation stipulates that if a human is ultimately responsible for a decision, the AI system may not be classified as high-impact.
While the government’s intention is to foster a supportive environment for AI innovation, the penalties for non-compliance—up to 30 million won (approximately £15,000)—and the challenges of adhering to the new regulations have raised alarms within the industry. A grace period of at least one year before penalties are enforced has been promised, but the uncertainty surrounding compliance remains a key concern for many.
Industry Response: A Mixed Reaction
The reaction from South Korean tech startups has been largely negative, with many arguing that the regulations are overly burdensome. A survey conducted by the Startup Alliance revealed that a staggering 98% of AI startups felt ill-equipped to meet the new compliance requirements. Lim Jung-wook, co-head of the alliance, expressed widespread frustration, stating, “There’s a bit of resentment. Why do we have to be the first to do this?”
Critics have also pointed to potential disparities in compliance, noting that while all domestic companies will face regulatory scrutiny, only foreign firms meeting certain thresholds—such as Google and OpenAI—will be subject to the same level of oversight. This could create an uneven playing field, further complicating the landscape for local innovators.
Civil Society’s Concerns
On the other side of the debate, civil society groups have expressed disappointment that the new laws do not go far enough in protecting citizens from the risks associated with AI technologies. South Korea has been identified as a significant hub for deepfake pornography, accounting for over half of all global victims, according to a 2023 report. Amid escalating concerns over AI-generated sexual imagery, particularly following a scandal involving the creation of illicit content on platforms like Telegram, the urgency for robust protective measures has intensified.
Four human rights organisations, including Minbyun, have voiced their concerns about the legislation, arguing that it fails to adequately safeguard individuals adversely affected by AI systems. They contend that while the act mentions protections for “users,” it primarily addresses institutions such as hospitals and financial entities, rather than individuals. Furthermore, the law lacks definitive prohibitions against harmful AI applications, leaving significant regulatory gaps.
A Different Path in AI Governance
In crafting its legislation, South Korea has chosen a distinctive approach compared to other jurisdictions. Unlike the European Union’s stringent risk-based regulatory framework, or the more market-driven strategies of the US and UK, South Korea has adopted a flexible, principles-based model. This approach, described by law professor Melissa Hyesun Yoon as “trust-based promotion and regulation,” aims to encourage innovation while maintaining oversight.
The Ministry of Science and ICT has expressed confidence that the new law will eliminate legal ambiguities and promote a “healthy and safe domestic AI ecosystem.” Ongoing revisions and clarifications to the rules are anticipated as the implementation process unfolds.
Why it Matters
South Korea’s move to regulate AI represents a significant milestone in the global discourse surrounding technology governance. As the nation strives to balance innovation with responsibility, the implications of its legislation could reverberate far beyond its borders. This pioneering effort may influence how other countries approach AI regulation, shaping the future landscape of technology governance on a global scale. As the world watches, the outcomes of South Korea’s regulatory experiment will be crucial in determining the trajectory of AI development and its societal impacts.
