[Column] How Not to Be Scammed by Chat GPT

[Asia Economy Reporter Yuri Choi] A student who learned poetry creation in a literature class wrote a poem. Whose copyright does the poem belong to?the teacher who taught the creation or the student? Naturally, it belongs to the student. However, when applied to artificial intelligence (AI), this question becomes less straightforward.

A good example is 'GitHub Copilot,' created by Microsoft (MS) subsidiary GitHub and OpenAI, the developer of Chat GPT. GitHub Copilot is an AI that generates program code. It learns from open-source code shared by developers to write code. It learns from humans and produces AI’s own creative works. Open-source developers filed a class-action lawsuit claiming that the AI illegally copied their code. MS submitted a brief asking the court to dismiss the lawsuit, stating that the AI did not cause any harm to the developers.

The question of whether AI can be held accountable for copyright is not simple. Various controversies are intertwined. It starts with whether the results created by AI can be considered creative works and whether the AI’s effort and pain in creation should be recognized. It extends to whether AI can be held responsible if problems arise with AI-created works, and if not, who should bear the responsibility.

The performance of AI implemented through Chat GPT is advancing at an incredible speed. However, the legal and ethical controversies surrounding AI are not progressing as quickly. Unresolved controversies appear in different forms. The U.S. online media BuzzFeed saw its stock price soar after deciding to use Chat GPT for content development. On the other hand, IT specialist media CNET retracted articles written by AI, citing many cases of plagiarism and factual errors. TikTok acquired AI startup Jukedeck, which automatically composes music for videos, producing over one million songs. Conversely, Google developed a similar AI but decided not to release it due to a 1% possibility of plagiarism.

Chat GPT is like an arrow released from a bowstring. Within two months, it has created an enormous syndrome, generating both expectations and fears. To prevent that arrow from turning against ourselves, we must answer the questions AI poses. When asked whether AI can be held responsible if problems arise with AI-created works, Chat GPT replied, "There may be controversy, but unlike humans, AI does not make decisions or have intentions," and "I think AI cannot be held responsible." Learning how to use AI effectively may be as important as learning how not to be deceived by AI.