South Korea’s AI Basic Act: five lessons for cities building trustworthy AI

The rapid acceleration of artificial intelligence has moved into the heart of municipal governance, with city leaders no longer asking if they should adopt AI, but how they can do so responsibly while maintaining a competitive edge.

For cities at the early stages of this journey, the path forward can feel daunting – caught between the desire for rapid innovation and the necessity of protecting residents’ interests.

Around the world, central governments have been forced to react quickly to the advent of every-day AI use and what it means for people, privacy and economy. Notably, South Korea, a nation that has consistently punched above its weight in the global tech arena, has in recent months passed its Basic Act on the Development of Artificial Intelligence and the Establishment of a Foundation for Trustworthiness (commonly known as the AI Basic Act). With the act’s passage through parliament, South Korea is codifying a comprehensive national strategy. For city planners, mayors, and digital transformation officers, the South Korean model offers vital lessons in balancing industrial growth with ethical guardrails. 

The South Korean context: a dual mandate

Promulgated in early 2025 and having taken full effect in January 2026, the AI Basic Act is distinct because of its "dual-track" philosophy. Unlike the European Union’s AI Act, which is often viewed as more regulatory-heavy, or the United States’ more decentraliesd, sector-specific approach, South Korea’s Act explicitly seeks to promote the AI industry while simultaneously institutionalising "trustworthy AI." 

For a city at the start of its AI journey, this duality is the most critical starting point. Too much regulation can stifle the startups that drive local economies; too little can lead to public backlash when an automated hiring tool or a predictive policing algorithm fails. 

Lesson one: the "control tower" model for governance 

One of the most significant features of the South Korean Act is the establishment of the National AI Committee, chaired directly by the President. This "control tower" ensures that AI policy is not siloed within a single technology department but is treated as a cross-cutting national priority. 

Early-stage AI cities often fragment their AI efforts across departments, each running their own pilots. South Korea teaches us the value of a centralised “digital command centre" at the municipal level. By creating a unified body that reports directly to the mayor or city manager, cities can ensure that data standards, ethical guidelines, and procurement rules are consistent across all urban services. 

Lesson two: defining "high-impact" AI 

South Korea does not treat all AI equally. The Act introduces a specific category called "high-impact AI." This includes AI used in critical sectors like healthcare, energy supply, transportation, hiring, and criminal investigations. Operators of high-impact AI face stricter requirements, including: 

• Prior assessment: evaluating whether the AI qualifies as high impact before deployment 
• Explainability: providing meaningful explanations of the AI’s outcomes and the criteria used 
• Human oversight: ensuring a "human-in-the-loop" mechanism is always available. 

Cities are the primary theatres for high-impact AI. Whether it is a traffic management system or an automated social services eligibility tool, the stakes are high. Emerging AI cities should adopt South Korea’s risk-based classification. Instead of creating a one-size-fits-all regulation that burdens a local developer making a simple chatbot, cities should focus their regulatory scrutiny on applications that directly affect the safety and fundamental rights of their residents. 

Lesson three: transparency as a non-negotiable baseline 

A cornerstone of the South Korean Act is the transparency requirement. AI operators providing content (sound, images, or video) that could be mistaken for human-created must provide clear notice. Furthermore, users must be informed in advance if a product or service is powered by AI. 

Public trust is the most fragile resource in any smart city initiative. South Korea demonstrates that transparency shouldn’t be an afterthought. Cities should mandate that any AI-driven public service – from AI-powered waste management notifications to automated customer service lines – be clearly labelled as such. Implementing "watermarking" for AI-generated government communications can prevent misinformation and build a culture of accountability from day one. 

Lesson four: investing in "trust infrastructure" 

The AI Basic Act isn’t just about rules; it’s about resources. It mandates government support for AI clusters, data centres, and R&D. Notably, it creates the AI Safety Research Institute, a specialised body tasked with evaluating risks and developing safety standards. 

The lesson for cities is that you cannot regulate what you do not understand. For a city to become a leader in AI, it must invest in "trust infrastructure." This means more than just building server farms; it means creating local "AI sandboxes" where startups can test their tools in a controlled urban environment under the supervision of city-funded safety experts. By providing the tools for safety – such as open-source auditing datasets – cities can attract responsible companies. 

Lesson five: the importance of phased implementation 

South Korea allowed for a significant preparation period (one year) between the passing of the its Basic Act and its enforcement. During this time, the government worked with lower statute alignment bureaus and expert task forces to draft the specific guidelines that would fill in the legislative gaps. 

AI innovation is moving faster than the law, and always will, so a rigid, permanent regulation passed today will be obsolete in the near future. To avoid this, cities can start with a framework policy that sets the vision and values, followed by a period of guidelines, pilot programmes, and ethics codes, before moving to hard enforcement. This allows the city to learn from real-world data and adjust its rules as the technology evolves. 

Strategic recommendations for city leaders 

For cities looking to implement these lessons, the following steps are recommended: 

• Draft a "city AI charter: define the core values (e.g. fairness, privacy, sustainability) that will guide all municipal AI projects 
• Conduct an AI audit of existing services: identify where AI is already being used in the city and classify those uses by risk level 
• Establish a local AI advisory board: consider including representatives from local universities, civil society, the tech industry and the local economy 
• Prioritise data sovereignty: ensure that the data generated by the city remains a public asset, used to train AI models that benefit the community rather than just private interests. 

The South Korean AI Basic Act serves as a reminder that AI governance is not just a technical challenge, but a political and social one. By integrating industrial promotion with strict safety protocols, the government is attempting to build an AI society that is both prosperous and secure. 

For cities still in the early stages of their AI adoption, the message is clear: it’s not a choice between being pro-innovation and pro-safety – these two goals are mutually reinforcing. By establishing a clear "control tower," prioritising transparency, and supporting local innovation, cities can ensure that AI becomes a tool for empowerment rather than a source of exclusion. The Basic Act can act as a helpful foundation for local governments; it is then up to the world’s cities, and the technology ecosystem that supports them, to build upon it, tailoring these lessons to the unique needs of their own communities.