I’m an AI expert – there are three scenarios where bots could go rogue & treat humans like ‘insects’
IF artificial intelligence goes wrong, there are three possible outcomes that each spell doom for mankind, an AI expert has exclusively revealed.
Mo Gawdat, arguably the most knowledgeable AI expert on the planet, told The U.S. Sun that as artificial intelligence increases, the “competitive nature of humanity” will also increase.
From Meta to Microsoft, the United States to China, the AI arms race is full of heavy hitters with dreams of world domination.
“It is almost inevitable,” said Gawdat, “that sooner or later, we will have a lot of decisions that are important for humanity – including some defense decisions – handed over to AI.”
And if AI goes rogue, there are three scenarios that are possible.
The first one is what Gawdat calls a “scenario of mistakes.”
In short, AI could make decisions that are aligned with its original objectives as assigned. These objectives would not align with the best interests of humanity at large. The results of such a misalignment would likely prove to be catastrophic.
Twenty years ago, the AI guru Nick published a paper titled Ethical Issues in Advanced Artificial Intelligence.
In it, the controversial philosopher included a rather compelling thought experiment. Aptly titled “the paperclip maximizer,” the mental exercise demonstrated the existential risks posed by a sufficiently advanced AI.
Imagine an AI whose sole goal involves nothing more than making as many paper clips as possible.
As Bostrom noted, a sufficiently intelligent AI would quickly realize that humans pose a direct threat to its objective on three different counts.
First, humans could switch the AI off; second, humans could change their minds and change their goals; third, humans are made of atoms, and atoms can be turned into paper clips.
In short, AI could turn on humans – not out of malice, but out of a “desire” to complete the initial goal. Because of this “desire,” the entire human race could be converted into paper clips.
The second scenario is what Gawdat calls an “element of conflict.” It is eerily similar to the first scenario. “For example, imagine a conflict around energy,” the former chief business officer at Google said.
AI is powered by data, and data requires storage. AI, said the Egypt native, “might find that New York City is a very good place to have data centers and that humans are using that space in an unnecessary manner.”
To make room for more data centers, AI might make decisions that are not aligned with human interests but entirely aligned with its own interests. Humans would be removed from the equation through any number of various methods.
The third scenario is what Gawdat calls “a question of irrelevance.”
For example, AI could simply take charge of some major infrastructure network (like the country’s power grid) and show absolutely no concern for humans because it would be so advanced that humans would be deemed utterly irrelevant.
Sooner or later, we will have a lot of decisions that are important for humanity … handed over to AI
Mo Gawdat
In this case, humans would be treated in the same way that we currently treat insects, with no concern for our welfare and well-being. The harm done to humans by AI would simply be a “byproduct of the main objective,” said Gawdat.
It follows an admission from the founder of ChatGPT, Sam Altman, that his technology is planning to go into strange new places in the coming years.
Altman, CEO of OpenAI Inc, which runs ChatGPT, said that he believes it will reach superintelligence one day, adding that proper conditions will need to be in place when that happens.
Speaking at an event in Abu Dhabi in the UAE, he said: “When we develop superintelligence, we’re likely to make some decisions that public market investors would view very strangely.”
And he said that due to the power OpenAI has, he will want to stay in full control of it and not take the software public.
The rapid growth of AI programs such as ChatGPT has also raised fears that millions of jobs will be wiped out worldwide.
At least one content writer recently revealed his work would be replaced by ChatGPT – virtually killing his business in a matter of days.
“It wiped me out,” Eric Fein told the Washington Post.
And AI has also been manipulated for grotesque uses in recent months.
In one horrifying TikTok trend, AI image generators recreated child murder victims, including their voices, to create disturbing videos.
The videos often share gruesome details about how the kids were killed, and come without a content warning.
Experts at OpenAI and Google Deepmind have warned that artificial intelligence could even lead to the extinction of humanity.
In a statement published by the Center for AI Safety, it said: “Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war.”