The platform industry will update self-regulation principles for search and recommendation services in line with the spread of AI. It plans to reflect new risk factors such as AI model changes and supply interruptions, data processing and responsibility structures, and to continue checking implementation.

The Ministry of Science and ICT-supported platform self-regulatory body on Tuesday held a meeting at the Payto Hotel in Seoul attended by platform companies and advisory members, as well as OpenAI and Anthropic.

Key agenda items at the meeting included stability of AI model supply, service responsibility structures, transparency in data processing, and links to open standards.

The platform industry assessed that AI model providers' usage fee policies, version changes and supply interruptions could directly affect service operations. Participants agreed that a sustained framework for cooperation with AI companies is needed, as platforms would find it difficult to resolve technical problems on their own even if they arise.

Platforms and AI companies also shared difficulties such as model version changes and supply interruptions without prior notice. They agreed to raise predictability of changes in the service environment by discussing mid- to long-term roadmaps and clear notification standards.

They also discussed ways to support switching between AI models based on open standards and expand interoperability among domestic operators. Participants exchanged views on the areas each of platforms and AI operators can control when problems occur, and on procedures for judging responsibility and responding. In the data area, they discussed standards for AI processing of platform data and ways to indicate information sources. The two sides agreed to continue consultations in response to the expansion of AI services through the platform self-regulatory body.

The platform self-regulatory body was launched in August 2022 to strengthen platforms' social responsibilities and discuss self-regulation measures for innovation and coexistence. It currently includes 13 platforms and six associations and groups. The ministry supports operation of its Data and AI subcommittee and its innovation-sharing governance subcommittee.

The Data and AI subcommittee decided to update the self-regulation principles for enhancing transparency in search and recommendation services, drawn up in 2023, in line with changes in the technology environment such as expanded use of generative AI. It plans to inspect implementation by platform and publish a collection of implementation cases based on results from the past three years.

The innovation-sharing governance subcommittee plans to share implementation cases by company of the platform corporate governance principles drawn up last year and publish a best-practices collection.