Few people deny the prediction that artificial intelligence (AI) will significantly change the world going forward. However, opinions are divided even among experts on exactly how it will change things. Opinions are even divided on whether AI technology will ultimately change the world for the better (optimism) or for the worse (pessimism). For example, the renowned futurist Ray Kurzweil has argued that when AI develops to reach the singularity, surpassing human capabilities, energy concerns will disappear and humans will enjoy lives approaching immortality. On the other hand, the late genius physicist Stephen Hawking argued during his lifetime that "AI could be the worst event in the history of human civilization. Preparation is needed to prevent a dystopia of technological absolutism." Professor Geoffrey Hinton, who developed the deep learning algorithm and is called the "Godfather of AI," even went so far as to say, "It is not impossible for AI to wipe out humanity. Because of such concerns, I regret what I have done."
Yet the world is now pouring enormous resources into AI technology development, centered on U.S. Big Tech companies, because of AI's outstanding productivity and the economic value it generates. Korea has also made leaping to become "one of the world's top three AI powerhouses" a major national agenda, and is concentrating national capabilities by creating a presidential "National AI Strategy Committee" and allocating large-scale research and development (R&D) budgets. Thanks to such efforts, AI technology is developing by the day. About three and a half years ago, after OpenAI's ChatGPT demonstrated the ability to generate text, AI soon became able to utilize not only text but also images and videos, and after passing through a reasoning stage, has now entered the Agentic AI stage, where it acts autonomously. In other words, an era has arrived in which AI, without human intervention, plans on its own and uses tools to produce complete results.
As AI technology develops rapidly like this, cases that pessimists had worried about are actually beginning to emerge. One example is the recent U.S.-Iran war. In this war, the U.S. and Israel have combined AI technologies to collectively kill Iranian leaders in bombings and designate attack targets in Lebanon and Iran, maximizing the use of AI technology in warfare. AI technology has begun to be used in earnest for the destructive purpose of war before peaceful purposes. Moreover, in the early stages of the war, a girls' elementary school in Iran was destroyed in an airstrike, killing more than 100 children, a disaster presumed to have been caused by AI misjudging the attack target. In this way, the possibility of errors in AI technology is emerging as a major problem. Furthermore, there are research findings that AI, once outside human control, exhibits more belligerent and hardline attitudes. Not long ago, when a British research team conducted a wargame simulation using AI models, nuclear weapons were used in 20 out of 21 high-risk scenario confrontations. Because AI has no ethical awareness or fear of human extinction, it regards nuclear weapons simply as one of the efficient options for victory and used them with a 95% probability.
A program called "Claude Mythos" recently released by U.S. AI company Anthropic delivered another shock. As concerns spread among experts that a program designed to find software vulnerabilities could conversely operate as a very powerful "AI hacker," governments around the world, including the U.S. government, have begun discussing countermeasures. The Korean government also convened security officers from major IT companies for an emergency meeting, and the Financial Services Commission discussed countermeasures with the banking and insurance industries. In this way, the misuse of AI technology can threaten the safety of human civilization.
Of course, AI technology, if used well, will boost productivity and can further contribute to enhancing human welfare. Realistically, with companies around the world racing to develop AI technology, we cannot stand by with our hands behind our backs. However, since side effects from the misuse and abuse of AI technology are now beginning to appear, we must not be lax in preparing for them either.
