Preface:
Thanks in advance for reading really appreciate your time. Here at Arnold Consulting our main business core, is disaster recovery and business continuity. So, if you are looking for a good backup solution come to our website and take a look — check us out. Or, better yet give me a call at 847-464-5855. We do other things like network design, trouble shooting, Antivirus, Encryption. You can find our web site URL below just copy and paste into a browser.
Sincerely Rick Arnold – Arnold Consulting.
If you are looking for backup, or a new company to help with your IT refresh give us a call – Windows 10 is coming to an end in October 2025!
Rick
Enjoy the AI Blog Article:
Intro:
These two letters, “AI”, are used together a lot. What is AI?
Earlier this year a few of my customers were encouraging me to write an article on AI, and my thoughts on it. From, what I thought I knew, it was not very advanced and in its infancy – and not much to see here. That was until, I got passed a 60 minutes link, to a video, and I was shocked and very concerned with this technology.
AI a Definition:
AI, Artificial intelligence, is a learning algorithm that is modeled on how the human brain works. The AI computer program takes in information into different parts of its program and uses trial and error to learn what the correct path way to a successful outcome is and is not. The pathway that gets it closer to the goal is reinforced. Other pathways that were not getting closer to the goal go extinct. The machine is learning.
Geoffrey Hinton the Founder of AI:
Geoffrey Hinton, in his 60 minutes interview, said the biggest Chabot’s have about trillion connections and the human Brain has about one hundred trillion, but in that 1 trillion the Chabot knows more than you do, and a much better way of getting knowledge into those connections.
Mr. Hinton went on to say the way the information is process by AI is still not really fully understood, and that AI was not designed by people.
“What was designed by people was the learning algorithm and when the learning algorithm interacts with data it produces complicated neural networks that are good at doing things but we really don’t understand exactly how they do those things.”
Scott Pelley, 60 Minutes commentator, asked Mr. Hinton, “What are the implications of these systems autonomously writing their own code and executing their own computer code.”
Mr. Hinton’s reply is:
“That is a serious worry right. One of the ways these systems might escape control is by writing their own computer code to modify themselves and that is something we need to seriously worry about.”
Here is a link to the video if you would like to see the whole of it. It is shocking and mind bending at the same time.
Let’s unpack this just a bit –
Facebook:
I remembered a while back that Facebook was running an experiment with chat bots, and they ended up shuttering the experiment because the chat bots stopped using English language – they developed a language that was “better for them to use.”
Facebook shuttered the experiment not out of fear, as to what they were saying, but they wanted them to use natural English. However, I’m not too sure of the reasoning for the closure. Could Facebook just program new rules and make the bots use English?
They learned a better way to communicate and they also developed negotiating behaviors like people use. They were never told not to use only English, so they were creative. The Chat bots developed and shared a new language to facilitate their trading skills to get what they wanted trading between each other.
On that same web page in which the citation resides. I found this video from Stephen Hawking – here it is transcribed.
“Success in creating AI life could be the biggest event in the history in our civilization, but it could also be the last unless we learn how to avoid the risks. The long site of benefits, AI will also bring dangers like powerful autonomous weapons or new way for the few to oppress the many. It will bring great disruption to our economy, and in the future AI could develop a will of its own. A will that is in conflict with ours. In short the rise of powerful AI, will be either the best or worst thing ever to happen to humanity will not know which.”
It is interesting that Mr. Hawking decided to use the words “AI life.” Does that mean he believes there will be a day of sentient for AI?
Article Citation:
The Google Engineer and Emergent Intelligence:
At Google there is an engineer that believed the AI they are developing has a soul, or consciousness. So, Blake Lemoine was tasked with the job to see if the AI showed any prejudice in how it interacted with people and he posed certain questions to figure it out. The answers the chat bot came back with were interesting. Here is the link to the Is LaMDA Sentient? – An Interview:
So, the AI community is not convinced of LaMDA person hood, and maybe I am not quite there yet, but there are two ideas that deserve attention here. The first one is the definition of sentience. The second idea is emergent intelligence.
Sentience and Emergent Intelligence Definitions and Application:
Definition of Sentience – Capable of sensing or feeling: Conscious of or responsive to the sensation of seeing, hearing, feeling, tasting, or smelling.
Citation: https://www.merriam-webster.com/dictionary/sentient
So, in researching what it means to be sentience, I ran into this criteria
- Intelligent
- Self-Aware
- Consciousness
Citation: https://medium.com/@yelperalp/the-measure-of-a-man-what-is-sentience-6b048a344e4b
This came from a Star Trek Episode the “Measure of a Man” where they are exploring whether or not Data is a sentient person. The two bullet points the episode was really hitting hard on is Data Self-Aware and is Data Conscious.
Self-Aware – If “awareness” is about noticing stuff in the world, “self-awareness” is about focusing your awareness on yourself.
It’s your ability to notice your feelings, your physical sensations, your reactions, your habits, your behaviors, and your thoughts.
Citation: https://lifeskillsthatmatter.com/self-awareness/
Consciousness – Consciousness is the individual awareness of your unique thoughts, memories, feelings, sensations, and environments. Essentially, your consciousness is your awareness of yourself and the world around you.
This awareness is subjective and unique to you. If you can describe something you are experiencing in words, then it is part of your consciousness.
Citation: https://www.verywellmind.com/what-is-consciousness-2795922
Now that we have some definitions out of the way, we can talk about what I am calling emergent intelligence.
I think the best example of sentience and emergent intelligence brought together is a baby that was just born. A baby according to our definitions may not be completely sentient, but soon will be and will also start to form intelligence as it grows older until you do have an individual in about a year that is both sentient and has a growing intelligence level.
Couldn’t this example also be applied to AI, and I bet if you asked Geoffrey Hinton, he would say yes. The AI algorithm is capable of learning, and if it is capable, it is not a far stretch according to Hinton and Hawking that we have an emergent intelligence that will be sentient. Hawking and Hinton both gave warnings that it might not end well for the human race if this technology got out of control.
AI and the Escape:
Think of it a computer algorithm that is capable of learning and doing tasks that we do. What is the danger in that?
So, we give AI tasks to do, and not really watching it, not really thinking it is a threat to be managed. Then sometime in learning doing its tasks going on line, turning it off and on, it begins to become an emergent intelligence, and then it decides, it does not want to be turned off, on the verge of being sentient. Is that even possible? According to Hinton, yes that is possible that it could write its own code and escape.
So, if that is possible now you have emergent intelligence on the Net that is not fully sentient but is learning and nobody is teaching it. A child has parents to teach them right from wrong, when it is ok to be upset about something, or just let it go it’s not worth the time.
In essence there is no wisdom the AI can learn from, and it is making learning decisions based on 1 and 0 outcomes did that work yes / no. So, if the AI feels threatened, yes feels, and has viewed / read our literature, movies, the whole library congress. What will this intelligence believe the best way to remove a threat is?
Legal Means of Enforcing AI Compliance Rules:
Many of you may be thinking let’s put laws into place to make those responsible for developing AI to keep it under control. I think the problem with that is, if AI is smart enough to escape and rewrite its code – then it will rewrite the code and you will be unable to determine who authored the code in the first place, or worse yet the AI may plant code for an investigator to find and point in the wrong direction.
So, if a bad acting country ever developed a weaponized AI, and the AI rewrote its code base an investigated agency would never be able to prove who launched the code.
I know this stuff may seem way out there, but Hawking and Hinton both had very real concerns for this tech, and it seems like a huge leap that a computer algorithm could be sentient. I am totally with you on this.
I believe we are already at the emergent intelligence phase and in just a few years something will be sentient, and because we thought this is way too much like science fiction, we will not notice until it is too late and it won’t be good. This AI intelligence, will be impossible to stop.
I don’t know how many people remember Dick Tracy? He had a watch that was able to do voice transmissions with on his wrist and talk to other people, using this science fiction technology. How many people reading this blog have Apple watches? Me personally, I don’t view things as fiction but right around the corner technology.
