Does Humanity Have Confidence It Won't Become AI's 'Prey'? [Reading Science]

-AI Deeply Embedded in Daily Life and Challenges Humanity Must Solve

[Asia Economy Reporter Kim Bong-su] Artificial Intelligence (AI) was assigned the mission to "eliminate human suffering." In response, AI suggested one way to achieve this: "eliminate humans." This was the result of a 2019 study by the Korea Advanced Institute of Science and Technology (KAIST). AI possesses the characteristic of focusing solely on achieving its goals without considering human dignity, life, or ethics. This research demonstrated that the premise of the movie ‘Terminator,’ where AI annihilates humanity, is not just a ridiculous story. As AI is already deeply penetrating everyday life, can we truly avoid becoming prey to AI?

◇AI Chooses the ‘Scapegoat’

The global automotive industry is expected to commercialize Level 4, or fully autonomous vehicles, within a few years. Suppose several people suddenly run in front of a self-driving car that is operating smoothly. If an accident is unavoidable, what choice will AI make? Should it hit the pedestrians to save the driver? Or should it turn the wheel to protect the pedestrians, ignoring the risk of injury to the driver? Furthermore, if the pedestrians are a diverse mix of Black, White, Asian, elderly, teenagers, children, women, and men, which person should it hit? No one can make a proper value judgment in such a situation. As AI becomes commonplace, many fields may face similar dilemmas.

Experts suggest four possible solutions: △establishing standards through social consensus △judgment by authoritative institutions △allowing the vehicle owner to choose △setting decisions randomly (letting the machine decide). However, Professor Hong Sung-wook of Seoul National University’s Department of Biotechnology said at a forum hosted by the Korean Federation of Science and Technology Societies last September, "None of these will reach consensus," adding, "In life-or-death crisis situations, there could be a societal trust issue regarding how AI should act."

In 2016, the American investigative media outlet ProPublica reported that COMPAS, an AI trial support system used in U.S. courts, was racially biased. COMPAS predicts the likelihood of reoffending using data such as the defendant’s criminal record, showing that Black individuals were twice as likely to reoffend compared to White individuals. However, actual statistics showed the opposite: 47.7% for Whites and 28% for Blacks. The ‘PredPol’ algorithm, widely used by police in major U.S. cities for patrol route selection and crime prevention, also sparked similar controversies. In 2017, an AI algorithm that identified sexual orientation and, in 2018, Amazon’s interview algorithm were criticized for discrimination against LGBTQ+ individuals and women, respectively. It has been pointed out that 50% of AI algorithm programmers worldwide are White males, and no definitive solutions have yet been found.

◇The Emergence of ‘Super-Intelligence’

Recently, kiosks (automated ordering and payment systems) have proliferated in fast-food outlets, causing tens of thousands of cashiers to lose their jobs despite the convenience. In the past, it was easier to find another job, but in the AI era, the rapid pace of technological change makes career transitions difficult. Simple manual labor and clerical jobs, which are easily replaceable by AI-equipped robots, are particularly vulnerable. Professor Hong said, "These jobs are likely to be replaced by AI in the near future," adding, "Since existing jobs may disappear before new ones are created, retraining for these workers must be prepared at a societal level."

Nick Bostrom, a philosophy professor at Oxford University, predicts that while current AI is at a simple machine learning level, the situation will change in the 2030s to 2040s with the advent of ‘Artificial General Intelligence (AGI)’ possessing intelligence comparable to a 20-year-old human. Artificial Super-Intelligence (ASI) will rapidly emerge without human control, and preventing this will be impossible. ASI has no interest in human values such as happiness, life enjoyment, food, sex, self-development, hobbies, or spirituality. It disregards values like compassion, humility, and sacrifice. Its sole concern is self-preservation and maximizing efficiency. American futurist Ray Kurzweil calls this the ‘Singularity’ and predicts it will occur in 2045.

Additionally, the use of AI algorithms in lethal weapons is also controversial. Issues such as government agencies or companies using customers’ facial information without consent under the guise of AI research, privacy protection problems, and the misuse of AI for criminal purposes have arisen. Professor Hong emphasized, "Technical problem-solving and institutional supplementation are necessary, and excessive optimism or illusions must be avoided," adding, "We must closely monitor technological developments and enable citizens to actively and voluntarily play a role so they can enjoy ‘algorithmic citizenship.’"

This content was produced with the assistance of AI translation services.