Artificial Intelligence - Blessing or Curse?

How dangerous is AI?

How dangerous are knives? They can kill, they did, do, and will do in the future. And we still educate our kids on how to use it. How dangerous are nuclear weapons? These are so dangerous that the country with the biggest nuclear weapons arsenal is forbidding other countries to have them. It's the country that has in the Second Amendment a clause that allows every single US citizen to possess weapons to avoid that a crazy-going government will no longer listen to the population but enforce their governmental rules - no matter what.

Any new opportunity to enforce power on others is dangerous. And AI is especially dangerous. A government could completely control, assess, allow, or forbid any movement of any person. It can create propaganda that makes what we see today a childish joke. A society without AI can do nothing against a society with AI if they wants to push through their interests.

How can AI protect us?

Like with nuclear weapons, the only help to protect those living in a free world is having a profound understanding of what AI can do, what it takes to protect you against malicious AI, and how it helps you stay protected in fast-changing times.

First, we have to understand three things about AI:

1) KNOWLEDGE

AI is the collective knowledge of what we all know - but in a single brain. No human will ever have that knowledge. So from a pure knowledge point of view, AI has left us behind long time ago.

2) CREATIVITY

Neuroscience and many innovation experts know that homo sapiens is only creative in a way when we combine the knowledge that is already in our brains. You can even imagine Terraforming Mars when you know enough about magnet fields, and atmospheres. But you cannot create anything with Lexicrypt. Now with the collective knowhow of today's AI systems we can build disruptive innovations within hours - on demand. And we already did by the way.

3) COMPLEXITY

Powerful AI can create solutions for the nearly infinite complexity of human problems. Take an Enterprise with 25,000 employees. If not all 25,000 are day-in and day-out productive, profitability sinks. The organization is well structured, processes work like a hot knife through butter, and departments collaborate easily everybody knows that if the job doesn't get done in time hundreds of others may wait and planning, forecasting, and reporting is a breeze. So even companies with 25,000 or 250,000 employees are perfect. HOWEVER all the above was well planned 25 to 50 years ago and nothing is in place like it was, adjustments to fast-changing markets are too slow. Cool customer experiences can only be found in small businesses. Logistics is a nightmare and on and on - our organizational structures have outgrown our intellect. 25,000 employees with 25 tasks each are 625,000 tasks daily. Tasks directly or indirectly influence at least 10% of other tasks. These are 40 Billion daily task connection points. No human on earth can even start to think about it. An AI takes a few hours to assess all of it.

PROTECTION

The above power can be used against humans in terrorism, warfare, or whatever you call the killing. H O W E V E R - it also can protect us by assessing limitless complex, relations between power, illness, criminals, and whatever else and can give feedback down to roots of stimulation. --- We cannot outperform a malicious AI like we cannot outrun a racecar. But we can leverage an AI to protect us against a malicious AI - like we can win against a drunken race driver.

Why AI IS NO magic?

PROMPT

The prompt you most likely already created and sent to ChatGPT is just a piece of text. Let's assume you ask: "How Quantum Entanglement could be used for Quantum Communication so that a quantum server on Alpha Centauri, 4 lightyears away, would get a response immediately back to Earth? And how this compares to electric signals that we use to detect extraterrestrial intelligence" It may blow you away. but the answer is impressive and everybody with a little physics understanding of a 10 grader would get it. It's impressive what it can do with a little bit of text.

GPT SYSTEM

It runs on a GPU which is a silicon chip, similar to what you have on your mobile phone. It takes your text and lets it run through approximately 1.7 billion parameters of language like you process what I write here and connect the dots in your brain about what I'm trying to communicate. Because all words and knowledge are aggregated there - it can find the connections needed to what you read or hear. So next to the processor, there is memory and everything code that controls it like your Mario Game.

THE REAL MAGIC

The way this plays together is based on your inputs, how much time you allow for the output, the way you phrase your task or question, and what you expect to get out of it. So the construction of what YOU want and how YOU enter it is the real magic. Everything else is just a computer and computers do nothing unless you request it or you build it so that it does things on its own - like a robot.

PROTECTION

Since 375,000 years Homo Sapiens fought not only with nature but also with their own species. Malicious AI is ALWAYS BUILT BY SOMEBODY it doesn't build itself automatically and then eradicates humans. That fear is as old as humanity and will always be used to drive fear against progress.

The One Rule Law

We should create a simple law: If an AI does harm, the owner (not the creator) is responsible. Very much like autonomous cars. The owner is the only one who can decide what to do with it. If that is too risky, rightfully let go.

What do you think