South Korea AI Basic Act

Chapter 4: Ensuring AI ethics and reliability (duties on AI business operators) – South Korea AI Basic Act

South Korea AI Basic Act 34.1.2: High-impact AI: plan for explaining outputs, their key criteria and the training data

What does 34.1.2 require?

So far as technically feasible, the operator establishes and implements a plan for explaining AI-generated outputs, including the main criteria behind them and a summary of the training data used to develop and operate the AI.

Maintained by Gerard Blokdyk

Other controls in Chapter 4: Ensuring AI ethics and reliability (duties on AI business operators) – South Korea AI Basic Act

28 Private autonomous AI ethics committee (voluntary)
30.3 Strive to obtain verification and certification before providing high-impact AI
31.1 Advance notice that a high-impact or generative AI product or service is AI-based
31.2 Advance notice that generative AI output is AI-generated
31.3 Label synthetic sound, image and video that could pass for real
32 Safety measures and reporting for AI trained above the computing threshold
33 Review in advance whether the AI is high-impact, and request confirmation where needed
34.1.1 High-impact AI: establish and operate a risk management plan

Query this from an agent

The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.

MCP endpoint All South Korea AI Basic Act controls