Novelist Hwang Sok-yong has acknowledged using ChatGPT in preparing his new novel, a disclosure that has intensified debate in South Korea about how artificial intelligence should be allowed to shape creative work. According to coverage of his return to fiction, Halmae, Hwang’s first novel in five years, revolves around a 600-year-old hackberry tree and spans centuries of Korean history, a scope he said he supported by consulting AI as an aid. (Sources: Changbi’s publication note and profiles of the book). [2]
Speaking on a YouTube broadcast, Hwang said, "I used ChatGPT as an assistant," and added, "I input about five or six elements, such as a 600-year-old zelkova tree, the historical backdrop, and composition methods, and had conversations with the AI." He characterised the exchanges as providing "the underlying framework of the novel." His openness reflects a widening practice among established writers who treat generative tools as part of their research or drafting process. (Background on the novel’s themes and publication details support this context). [2]
Yet the creative industries are moving in several directions at once. Major publishers and competition organisers have begun to close ranks against unlabelled AI involvement, imposing explicit bans or revocation clauses in prize rules to protect what they regard as human authorship. Industry moves elsewhere in the cultural sector echo this caution: leading webtoon platforms recently told entrants they must not rely on AI tools, requiring proof that scripts were human-generated. These actions illustrate a wider effort to draw boundaries around automated assistance. [3],[5]
Publishers and contest organisers admit enforcement is problematic. While some houses now state that works created with AI are ineligible or that awards may be rescinded if generative tools were used, executives say there is no reliable, non-intrusive method to verify whether an unpublished manuscript was aided by AI. That gap leaves editors weighing stylistic intuition and claimant declarations rather than technical detection, and it raises contractual questions about authors’ rights and the handling of manuscripts in review processes. [5],[6]
Others argue for a different path: instead of attempting to exclude AI entirely, the literary world might acknowledge collaboration by creating distinct categories for human-only and AI-assisted works. Proponents say such a framework would allow experimentation without eroding standards, while critics worry it could normalise outsourcing of originality. Meanwhile, state-supported initiatives show a contrasting posture; a government-funded communications institute has launched a contest specifically inviting AI-created imagery to promote Korea, demonstrating that policy and promotional programmes may take a more permissive approach to creative AI. [4],[3]
Policy developments are beginning to provide legal contours for these disputes. South Korean guidelines now allow copyright registration for works where a human has made a creative contribution using generative AI, while excluding outputs produced entirely by machines. At the same time, a draft enforcement decree for the country’s AI framework includes mandatory labelling requirements intended to make it easier for consumers to identify AI-generated material, though experts caution that practical verification will remain difficult. These measures aim to balance protection for human creators with recognition of tools that legitimately assist them. [5],[6]
As Hwang’s public comments show, prominent creators are testing the possibilities of AI even as institutions scramble to update rules and regulators try to set standards. The experience of other sectors , notably education, where an early roll-out of AI-powered textbooks encountered resistance and setbacks , underscores the complexity of integrating new technologies into established cultural practices. The debate now unfolding in publishing will likely shape not only contest rules and contracts but broader definitions of authorship in an era when human imagination and algorithmic assistance increasingly intersect. [7],[6]
Source Reference Map
Inspired by headline at: [1]
Sources by paragraph:
- Paragraph 1: [2]
- Paragraph 2: [2]
- Paragraph 3: [3], [5]
- Paragraph 4: [5], [6]
- Paragraph 5: [4], [3]
- Paragraph 6: [5], [6]
- Paragraph 7: [7], [6]
Source: Noah Wire Services