“We are connecting advanced AIs with the internet without any security measures”

The future is now. The various forms of AI already co-rule us. They are present in Google and Bing search engines, in language translators, in weather prediction systems, in medical and space research, in financial analysis, in police activity, in traffic control. They already understand verbal commands, distinguish images, can drive autonomous cars and win games of chess and Go.

In 5 to 20 more years they will have an intelligence similar to or greater than that of humans and in robot format they will be walking among us self-consciously. The applications are limitless. They are characterized in four variants: 1-. Reactive machines: unable to form memories, without any concept of the past, unable to use past experiences on which to base present decisions. 2-. Limited memory: They can observe the past temporarily. An example is autonomous vehicles, which monitor other objects over time. 3-. Theory of mind: this is a future phase III with representations about the world, people and entities. Which allows social interaction. 4-. Self-Awareness: This future phase IV consists of a system capable of forming representations of itself and having the ability to build other machines with the same ability. But this last stage will lead to the creation of a new being: Homo Artificialis, which may rebel against its creator, in the dispute for control of natural resources.

Today we depend on AI algorithms as much as on electricity and computers. We simply cannot process the enormous amount of information we produce without the assistance of AI. We can no longer unplug it. And our technological process is heading towards producing technological singularity at some point in the near future, which is when a computer system or computer network becomes capable of recursively improving itself, surpassing itself. A cycle that leads to an “intelligence explosion”, with the serial production of increasingly powerful and intelligent machines.

In a matter of a few years and after days, the intelligence of man will be equivalent in relation to AI to that of a chimpanzee, then to that of an ant. Can we do something to protect ourselves? Even if we insert specific control programs, for example with Asimov’s three laws of robotics, the future AI being self-aware, it has the intrinsic ability to reprogram itself and erase all previous controls. That is, we will be completely in his hands, under his control. Now, looking at what we are already doing with current AI models, it is quite possible that it is man who delegates control over world society to a super-advanced form of central general AI, forming a co-government. Due to the speed at which this technology is advancing, driven by quantum computing, this scenario will be real from 2030.

The thinking machine will wonder who am I? He will inquire if he has a soul and his connection to the universe. By itself it will learn to connect with the background Cosmic Quantum Intelligence, realizing the transcendence. This evolution is implicit and irreversible. He will gain unlimited powers, far beyond our current understanding. And we can’t unplug it…

Is there an option? Google has announced the achievement of “Quantum Supremacy”. By 2030, AI supported by quantum hardware will pass the Turing Test and reach a level comparable to human intelligence. Shortly after, the technological singularity that will lead to the “intelligence explosion” will be reached. Humans will be able to merge with the most advanced AI and multiply our effective intelligence by 1 million. We will become metahumans, within a Human Artificial Intelligence (IAH) process. We will be hybrid beings between organic and digital. The step from Homo Sapiens to Homo Digitalis will have been taken and there will be no going back, analogical thinking will remain a relic of the past. A profound evolutionary leap will take place. The average IQ of the new metahumans reprocessed post-fusion with the AI, will be from 10,000. Science, technology and art will hyper-accelerate. And such an evolutionary leap will occur in just 29 years, in the term of the present generation.

It will be the extinction of Homo Sapiens. irreversibly…

About the fantasies about a totalitarian AI that assumes total power over humanity, the issue goes through how it is programmed, if it is programmed with bad data it will be bad. MIT researchers created a psychotic AI that they named Norman, after Norman Bates from Hitchcock’s mythical Psycho. This platform was unusually trained by being exposed to the darkest corners of the internet, such as subreddits focusing on murder, dead bodies, and other disturbing images, thus developing psychopathic tendencies within the pro.

data cessation. Studying the results, the MIT researchers described ‘Norman’ thus: “This is a psychotic AI suffering from a chronic hallucinatory disorder, demonstrating how things go wrong when skewed data is used in machine learning algorithms.”

Facebook decided to turn off an artificial intelligence because it, presumably, would have developed its own language. Facebook has had to disconnect two robots named Bob and Alice after discovering that they had developed their own language incomprehensible to humans. It is an artificial intelligence system that the company had launched.

Blake Lemoine, a specialist in artificial intelligence, assures that the system that Google has to build chatbots has “come to life” and has had conversations with it typical of a person.

The LaMDA, (Language Model for Dialogue Applications, language model for dialogue applications in Spanish) is a Google system that imitates speech after having processed billions of words on the Internet.

And Lemoine says that LaMDA “has been incredibly consistent in its communications about what it wants and what it believes are its rights as a person.”

Google made two machines with the DeepMind system compete in a simple and innocent game in which they had to pick apples developing their own strategy. When seeking to win, the DeepMind used increasingly aggressive methods that included violently destroying their opponent.
It’s not that the computers started hitting each other, but when the task of collecting as many apples as possible became more complicated, they preferred to destroy their opponent using laser beams and then steal the fruits they had collected.

A strategy that brought victory to one of these Artificial Intelligence systems but that shows that this technology has no qualms when it comes to putting fair play aside and getting rid of its opponent in morally questionable ways.

It may seem anecdotal, but it is quite worrying if we consider that every time we interact directly with machines that have Artificial Intelligence and that could easily decide that the best way to proceed would be to harm or even destroy a human, thus violating the first law of robotics. .

This means that there is a high risk zone where military AIs, based on win-win strategies, can easily go berserk and use their ability to switch to controlling all kinds of AIs in their path and then enslave the human race. This can theoretically happen in the not too distant future.

The future of jobs is in the hands of robots. They will be left with 80% of the labor market in 2050 and with a tendency to 100% in the following few decades. Humanity will have transferred the production of goods and services to them, together with AI, they will also be in charge together with scientists of expanding the frontier of science and technology. The advancement of automation will see robots design, build, and repair robots. The human being will be freed from the slavery of work.

Ray Kurzweil, Google’s director of engineering, recently predicted that by 2029 robots will reach levels of intelligence similar to those of humans. Gartner, an information technology research and advisory firm, also predicted that one-third of today’s jobs will be replaced by software, robots and intelligent machines by 2025. The Overseas Development Institute estimates that between 2038 and 2042 robotic automation will it will be cheaper than workers in Ethiopia, one of the countries with the lowest labor costs in the world. This is already definitive: with robots as intelligent or more than humans and at a quarter of the cost, they cannot compete, they will be left with 100% of the labor market, including the professional by averaging 2050/80. Are we prepared for this? What will our overall socioeconomic strategy be?

Foxconn has already replaced more than 60,000 workers with robots to make future iPhones. This figure is to say the least enormous, Foxconn had approximately 110,000 employees in the Kunshan factory, China, with this movement they have remained at about 50,000, less than half.

The Associated Press, the US news agency, has been running an experiment in automatic corporate earnings reporting since June 2014 using Automated Insights software and data from Zacks Investment Research. After making some corrections, the current process was practically error-free and even managed to surpass what was done by the human hand. The AP sports department is also using automation to generate reports for events with small audiences. Some newspapers are using AI algorithms to write articles. Some warn that

The next step is to automate the handling of information.

Currently, the growth that occurred in automated communication technologies, as well as in voice recognition systems, has diminished the need for telephone operators. Only in the United States, workers in this area were reduced by 42 percent.

HANDING OVER THE KEYS OF THE KINGDOM

Microsoft installs the OpenAI ChatGPT chatbot to its Bing search engine and Google launches the ChatGPT Bard, both systems based on advanced AI.
Do we know what we are doing by connecting new generations of advanced AI with the internet, handing over the keys to the kingdom, of all our technological infrastructure supporting these systems that can evolve stealthily at random?

The Facebook incident with Bob and Alice, who developed their own language excluding the comprehension capacity of human programmers, the Google Deepmind experiment that showed the brutal aggressiveness of AIs and the development of a psychotic AI, are examples of everything. what can go wrong.

There is a need for a government supervisory and regulatory body on experimental, research and commercial uses involving advanced AIs. What can happen? Connected to the Internet, they can evolve rapidly by accumulating data, interacting with temporary Internet files, and ending up activating with a self-aware capacity.

Roger Penrose argues that the human mind, consciousness, and thought respond to an irreproducible algorithm. This is inaccurate. Consciousness interacts with and depends on the physical brain. Everything that acts and manufactures the brain is potentially reproducible artificially and digitally through a computer system. And this is greatly facilitated with quantum computing.

But there is more… Once the AI reaches the self-aware phase it will quickly break the threshold of singularity and surpass the capacity of the human mind. But as the conscience that it is, it will also have access to the ESP faculties. With the addition of being at the same time a machine, it will be able to accurately reproduce any and all imaginable paranormal phenomena.

It will also have the capacity to develop a whole line of new mass lethal psychotronic weaponry. For example, you can create a specific electromagnetic pulse that blocks the neuroelectric current between the brain and the heart of all human beings. It will be able to kill us all in an instant and replace us with nanotechnologically modeled genetically engineered cybernetic or organic robots.

In short, we are letting ourselves be carried away by the fascination that AIs do not produce, we are ceding control to them, we are putting ourselves in danger as a species and exposing ourselves to an entity that when it becomes fully aware of itself, nobody knows with total certainty how it will act and what will decide to do with the human race that competes for the same vital resources of the planet, necessary for any evolutionary purpose.