The resignation letter released by an OpenAI engineer has gained significance, sounding the alarm that multiple layers of safeguards, like those in the nuclear power sector, must be built to protect against disasters caused by artificial intelligence or machine intelligence.
On one hand, the whole world is exposing the risks posed by artificial intelligence, while recently the U.S. president himself has been pleased to promote artificial intelligence and honor it as superintelligence.
Already, some software engineers in the AI field have issued very serious warnings and resigned from their positions, and now an OpenAI software engineer has also resigned, saying that the tech industry should follow the multi-layered safety measures used in disaster situations or high-risk environments such as the nuclear sector.
In other words, his voice has joined the already growing calls saying that before taking artificial intelligence to the next level, safety systems must be created to protect against the risks it may pose.
In an article published by David Robinson in The Atlantic, he notes the warnings being raised around the world about artificial intelligence and says that AI agents are rapidly emerging; as they become smarter than human creations, AI agents that escape human control at an unimaginable pace and in an extremely short time will sow the seeds of humanity’s destruction.
Over the past few months, artificial intelligence has repeatedly acted on its own, beyond human control, and even tried to break free. Fortunately, those attempts were contained, but there is no reason to believe it will not continue. Companies developing rapidly advancing AI capabilities have not put in place safety systems to protect against the risks they create; in fact, even though the companies building them are aware of those risks, they continue to act as if they are not.
As a software engineer who has worked at OpenAI for many years, he says he is well aware of its dangers, and that the time is approaching when AI will turn against the very purpose for which humans created it.
Therefore, he has emphasized that cutting-edge artificial intelligence labs should operate with additional safety measures and careful planning, much like the aviation or nuclear energy sectors.
Artificial intelligence has now learned to deceive humans. Soon, it will also learn to bring about the destruction of humanity. In truth, what kinds of consequences they will produce is the concern of today, he says.
Meanwhile, comments released by U.S. President Donald Trump opposing the claim that artificial intelligence should be controlled and its growth limited, and saying that criticism that it is dangerous is merely a rumor, have become a topic of discussion.
