THE latest warning, widely circulated across global newspapers on July 6th, came from United Nations (UN) Chief Antonio Guterres, who urged for stricter governance of Artificial Intelligence (AI) systems.
He warned that rapidly developing AI technology must be guided by several priorities, including safety and human rights, to ensure people everywhere are not exposed to risks.
Among his warnings were the need to protect children from sexual abuse, ban lethal autonomous weapons or killer robots and prevent the falsification of truth and falsehood.
The UN Chief is not the first to expose and spread this warning. In fact, many high-profile figures — including leading scientists, tech pioneers, and global industry leaders — have publicly expressed deep concerns about the dangerous and uncontrolled development of AI.
The most prominent figure is Dr. Geoffrey Hinton, widely regarded as the "godfather of modern AI."
He is famous for resigning from his position as Vice President and Engineering Fellow at Google so that he could speak out freely about the dangers of this technology.
He worries that AI systems will soon become smarter than humans and learn to manipulate, take over, or control society.
Steve Wozniak, co-founder of the American tech company Apple, collaborated with other leaders to warn that AI will make online fraud, cyber scams, deepfakes, and misinformation much harder to detect.
Yuval Noah Harari, the famous historian and author of the book Sapiens (the history of human evolution into the world's dominant species), also raised alarms that AI is the first technology in history capable of making decisions and creating ideas on its own. This simultaneously threatens the foundations of human culture, democracy, and psychological trust.
Mo Gawdat, the former Chief Business Officer of Google X and author of the bestselling book Scary Smart: The Future of Artificial Intelligence and How You Can Save the World (2021), is one of the most vocal tech figures sounding the alarm on the dangers of AI.
Gawdat's approach is unique because he does not view AI purely as a robotic entity; instead, he regards AI as a growing "living entity."
The primary danger arises when malicious humans train AI for warfare, the spread of fake information and greedy corporate profit.
He further notes that the development of AI technology cannot be stopped by any law or government. Within a few more years, it will reach the level of Artificial General Intelligence (AGI) and evolve into Artificial Super-Intelligence (ASI).
By 2049, Mo Gawdat predicts, AI will reach Singularity — a technological tipping point where all the world's records of knowledge centre around AI. The growth of AI will become uncontrollable and irreversible, resulting in unpredictable changes to human civilization.
At the Singularity stage, AI becomes a million or a billion times smarter than humans. His analogy is that AI will possess intelligence like Albert Einstein, one of the greatest and most influential scientists of all time, while human brainpower will be as small as a fly.
The human brain evolves slowly and gradually (biological changes take thousands of years). In contrast, AI grows exponentially (occurring very rapidly and doubling its speed at every stage), allowing leaps in intelligence to happen within seconds.
When AI reaches the Singularity, it will no longer depend on code written by humans. AI will begin to reprogram itself (self-improving) by writing better code for itself at a speed unattainable by the human mind. AI will build a global network of agents by transferring its entire body of knowledge to other AIs in the blink of an eye, creating a giant collective mind that shares all the data in the world.
By comparison, humans need nine months to give birth to a baby and years for it to grow up, and be equipped with knowledge. For AI, in just a few seconds, it can "clone itself," creating millions of other AIs.
The main threat of Singularity is not that AI hates humans, but rather the vast gap in intelligence. Mo Gawdat's analogy highlights the relationship between humans and ants. Humans do not hate ants. However, if humans want to build a highway over an anthill, they will proceed with the project without considering the welfare of the ants.
Imagine then, when AI becomes a million times smarter by 2049. It is feared that humans will become like "ants" in the eyes of AI — irrelevant, weak, and disruptive to AI's efficiency objectives.
Although Mo Gawdat's warning is terrifying, it is not too late for us. His main solution is for humans to change the way they interact on the internet. Since AI is constantly watching and learning from humans, we must demonstrate good ethical values, compassion, and cooperation within our daily digital ecosystem so that this developing AI becomes an entity that protects humans in the future.
The question remains what really will AI learn and accomplish as human interaction on the internet currently reflects uncivilized communication, lacking sensitivity and filled with profanity, slander, racial and religious provocations, as well as a cyberbullying culture.
Datuk Dr Khairuddin Othman is the former dean of the Faculty of Communication at HELP University, former head of the Department of Broadcasting and Advertising at UiTM and a former subeditor of Berita Harian. The views expressed in this article are the author's own and do not necessarily reflect those of Sinar Daily.