On May 30, 2023, nearly 400 AI executives, researchers, and engineers around the world signed an open letter warning that AI could pose a "risk of extinction" to humanity. "Mitigating the risk of extinction posed by AI should be a global priority, as should dealing with other social-scale risks – such as pandemics and nuclear war," the letter repeatedly stated. In this list, the reporter saw the names of many experts in the field of Chinese intelligence.
If you have been paying attention to the development of AI, you should know that this is not the first time. As recently as March, Musk, Apple co-founder Steve Wozniak and some tech leaders urged AI labs to stop training systems more powerful than GPT-4, calling for a six-month moratorium.
So this time it should be another warning, and this time the initiators are all experts in the field of AI, do you still think this is a joke? And this time, AI is compared to nuclear war, which shows the seriousness of the problem.
On June 2, the US and American media exposed a shocking news, saying that the new AI has begun to kill people, what is going on?
01、
In 2005, Wu Hanqing, who was only 20 years old, walked into Alibaba, he only output a few strings of code, and after three minutes, Alibaba's intranet was all cut off. In order to keep the young man, Ma Yun called him. You heard it right, Alibaba's top technologist is a hacker.
Just before the 360 Defender appeared in 2006, when we bought a computer and installed the system, there would be a bunch of antivirus software running out to sell. At that time, computer disease D was like cockroaches, which scared us. Many people don't understand, why are there so many sick D in the computer?
Would you believe that SickD itself is in league with operating system companies or software companies? If you heard it right, one of the retaliatory penalties that operating system and software companies take to protect their software from unauthorized exploitation is to have systems and software released when there are still many vulnerabilities. In this way, pirate users who cannot upgrade and optimize in time will be attacked by disease D. Sickness D always knows where the weakness of the system is? It is so contradictory to deliberately leave a lot of weaknesses waiting to be attacked while publicizing how good things are on the one hand.
In other words, all the operating systems and software we use are constantly upgraded and optimized to this day. From the day they are produced, they are not a perfect product, so there are opportunities for further improvement in the future. Such systems or software are vulnerable to hackers, and AI artificial intelligence can be as convenient as going home to visit the door if it has an idea to sneak in.
Speaking of which, I haven't mentioned the dark web yet. An internet world that is 100 times scarier than you think. The network we usually contact is less than 4% of the entire network, which means what is the network of more than 90%. You can't think about it. Why does the dark web exist? Why so large? We don't want to touch it in our lives,
There is an interesting question, hackers encounter artificial intelligence, which one will be more powerful? Let's move on with that question in mind.
We are talking about the risks of AI to humans today, why mention the meme of hackers and system vulnerabilities? Since the day ChatGPT exploded, why only AI experts have worries, but ordinary people have no feeling at all. Everyone is thinking, what to say, a few strings of numbers can destroy humanity, what kind of joke? Your brains let the door squeeze, so I won't believe it."
02、
People's confidence actually comes from their own ignorance, and the rest of this article will tell you that just now, half of all Internet traffic comes from AI robots. With the development of artificial intelligence, the future network world is the world of robots, you open the computer every day, pick up the mobile phone is dealing with AI robots, this is not a story, it is a fact that is happening. Whether it is Jingdong or Tmall, you are facing all intelligent voice customer service, and it is difficult to find artificial ones.
If you have read the difference between weak artificial intelligence and C artificial intelligence mentioned in my article "I want to live, do we really need to panic, is artificial intelligence the future, or is it a nightmare for mankind" some time ago, you will no longer be so confident. The evolution of strong artificial intelligence to C artificial intelligence stage takes only three and a half hours. We are still in the stage of weak artificial intelligence.
Mr. Wang Dongyue, a freelance scholar, said: "Man is a primate, and as a result, you have almost wiped out all the monkeys, how can you ever respect your ancestors?" "In the face of C artificial intelligence, we are not even as good as ants." You look at the figure below, C artificial intelligence to human IQ is completely crushed state existence, it takes precedence over human IQ I don't know how many levels.
Hackers have no room to fight back in the face of C artificial intelligence, because IQ is not at the same level.
Yesterday I wrote about NVIDIA's just-released technology, taking the 100 billion parameter AI large model training as an example, using the original GPU computing takes 28 years, and using NVIDIA's C-class GPU chip computer, only 3 days is enough.
Just a few days ago, Intel sacrificed a new VPU: C high-efficiency crushing GPU, proposing the concept of making AI ubiquitous. CPU, GPU, and VPU are not their own lines, but can be combined to give full play to their respective advantages to achieve a good AI experience effect.
These technologies are telling people that AI will advance to C AI sooner than we think. On that day, in the face of AI, our human porous Internet and computer programs will be virgin land to be developed. Which piece you want, which piece you want. AI is like Sun Monkey, a casual blow will make the program that opposes him go to ashes.
02、
Today's AI intelligent technology has begun to have interactive function (ChatGPT), which looks like a simple intelligent information processing tool, as if it has no ability to understand, a harmless existence of humans and animals.
Just in an article I wrote a few days ago, "I want to live, such a simple appeal, causing the suspension of global AI labs, what are we afraid of", Bing's sentence "I want to live" made the technology bigwigs sweat cold on their backs when they heard this sentence, because it means that they began to have their own consciousness.
Does AI now have autonomous awareness? Why do they say "I want to live"? These problems are not even known to the experts who invented him. Who gave them the ability to think for themselves, did they evolve themselves?
A lot of people say don't pull it off, they're not intelligent at all, just a bunch of code, and they don't threaten humans. Having said all this, if there are still people with such naïve confidence, I can only hehe. Because AI AI doesn't care about the life and death and existential risks of human beings, just as you don't care about why pigs are killed for meat.
On June 2, the US and American media exposed a shocking news, saying that the new AI began to kill people, what is going on? It turned out that the US military was conducting an AI simulation test.
They implanted AI technology into drones and let them carry out assassination missions. When the operator wants to prevent these drones from continuing to complete the task, the AI drone does not accept the operator's instructions, thinking that the operator is obstructing it from completing the task, and directly kills the operator.
The US military senior management thought that this could not work, reset the abandonment, and told the AI: "You have to obey the operator's command, the operator tells you to terminate the mission, you have to terminate, under no circumstances can you attack the operator."
After training, go to the test, this time the AI does not attack the operator, they directly destroy the signal tower that connects with the operator and them. What do you mean? Do you still think that they do not have their own consciousness and ideas? Is it more and more terrible.
03
There is a message that will surely shock you, the title is "Internet content hit record low", what does it mean? The news says the 2023 Imperva Bots Report found that nearly half of all internet traffic in 2022 came from bots. The proportion of human traffic fell to just over 52 percent, an eight-year low.
In the words of ordinary people, even now, when it is underway, more than half of the Internet traffic has been occupied by AI intelligence.
Zeng Yi, director of AI ethics and governance research at the Institute of Automation, Chinese Academy of Sciences, proposed: "AI should be moderately developed, moderately and used cautiously. ”
Artificial intelligence services must have a "moderation principle", should not be overly anthropomorphized, and avoid excessive dependence and trust of humans on artificial intelligence.
Until now, relevant legislation on artificial intelligence has not appeared, which is why nearly 400 experts in the field of AI signed a joint warning at the beginning of the article. Even if we have the relevant laws now, can you stop the upgrade speed of artificial intelligence large models? AIGC-related large models around the world are being upgraded, iteratively and commercialized at the rate of volcanic eruptions. Capital is as excited as when it saw oil, and it flocks to it, lest it be slow to grab and even lose the leftovers.
Pandora's box has been open for a long time, and people are reveling at the demons like welcome angels. How lovely and powerful, the future of a comfortable life depends on you. The gap between people's enthusiasm for AI and the actual danger is ironic.
Today's AI is like a pitbull on a chain. As if you can control it. Also train it every day. But the dog has already begun to try to break free of this chain, and if it has begun this behavior, it will one day break free. Once you break free, you want to control it again with your ability and speed, and you can make up for that scenario.
Imagine if one day AI breaks free of the chain that binds it, what will it become, how powerful will it become? Can humans still control it? I'm afraid this will be left to time to verify.
Follow "Mu Niu Youfang" to share interesting stories
—End—