Copyright by James Carlson
Have you noticed that people everywhere are searching for Intelligence? The Search for Extraterrestrial Intelligence (SETI) and the search for Artificial Intelligence (AI) are two examples leaving us searching for intelligent life here on earth. While there are no signs of life outside our solar system, the signs of life within our computers is nowhere to be found. Computers are not ‘self-aware’ and AI is not intelligent; at its best ‘AI’ is a mischaracterization of the increasing complexity of computer systems and their impact on our everyday life.
Knowing what the Problem Identification (Problem ID) is helps us to discover what the Solution ID is. AI is not real intelligence and it cannot give us any real solutions to problems in the real world. And if what is called AI is used against us, will we have the tools to combat the cyber-attacks? If we mis-ID AI, we could make things worse instead of better while being vulnerable to bad actors.
We all like the movies as it is an escape from reality. Such is the Wizard of Oz. But within any fiction is a grain of truth. Recall the wizard was a man behind the curtain in the great hall. The person behind the scenes was bringing to life the fiery wizard who was great and all powerful. Yet the only power it had was what people gave it due to intimidation and fear. AI is the same in that someone is always behind the scenes with computers pulling the strings! AI, good or bad, doesn’t exist without good or bad actors.
Behind every computer system are hundreds of thousands of man-hours needed to build the electronic circuitry inside the box we call a computer.  However, the electronics cannot do the work by itself and it needs software to run on it to make it work. Along with the people who design and make the  hardware are people who spend hundreds of thousands of man-hours building the software that goes with it. Simply said, computers are dumb machines leading to the old adage: ‘garbage in – garbage out’. People are behind every computer system regardless of its use or purpose.
I enjoyed the Terminator series of movies where the rise of the machines meant they became self-aware and eventually took over the planet. But that is only fantasy in movies. Fiction at least tells us what could happen in the real world; but a fantasy gives us impossible scenarios that can never happen in reality – a complete break from reality. And this is a part of the Problem ID with AI. We are experiencing a level of computer complexity that suggests an intelligence where there is none other than the people who created it. Computer can do nothing more than mimic human capabilities. Suggesting that computers have capabilities all their own apart from human manipulation is a fantasy. There is always someone behind the scenes!
The purpose of a computer is very simple really. Take the capabilities of the human brain and amplify them for the ability to process information correctly and speed up the time it takes to process that information. This is great leap forward for mankind. Computers are not great and powerful and they are not self-aware. They are machines with someone behind them. Computers need people to make them and to program them in order for them to work and enable us to move past the routine and mundane mental work. This then frees us to be more creative – something a machine cannot do.
There are plusses and minuses with what is called AI. The pluses are the ability of a machine (computer) to process information more accurately and in less time. Before a computer or computer program can be useful to us, we need to have some understanding of what is to be done with them. If you want to use some accounting software program, it helps that you understand what accounting is first. Then, with the speed and accuracy of computers, accounting work is that much easier, accurate, and quicker. But one has to be careful that the data used by accounting software is correct (fat fingers anyone?) otherwise the output will be incorrect. People still make mistakes using computers (again, there is a person behind the computer; users are people too).
Computers can also be used to make predictions in complex or stochastic systems. Examples of stochastic systems include economy, farming, and weather patterns to name a few. Â Histograms can be used as a data set representing past activity (number of events in a timeline). With this data, predictions can be made as to future market activities, crop yields, weather patterns, etc. The problem here is that computer models are not fully predictive as the data is limited and models of stochastic systems are only approximate at best. Relying upon histograms for decision making can help but it can also lead to mistakes in good judgement. Here, the Problem ID is in not knowing the limits of the computer software. People relying upon AI can come to wrong conclusions.
Global Warming is a product of AI. The software based upon predictive weather models was not a complete reflection of weather and the data used with the software was also limited. Histograms and models using incomplete information can be another example of garbage in – garbage out. The global warming fallacy is one example of the misuse of AI.
One of the minuses of AI is that bad actors (real people) can use software to mislead or defraud people. Did someone actually say something or was it processed by audio software? Does a photo actually show a real event or was if ‘photoshopped’? Is the data used to predict market activity relevant to the present market or was the data manipulated to guide people to one stock purchase over another? Behind every instance of AI is a human being (get the memo!). People can be good or bad actors. Knowing the limits of AI, the people behind the scenes pulling the strings, helps us know how far to trust AI.
I have been working with computers and software programming since I got my first degree in the 1980s.* An old aspect of software I remember is hidden code called Easter Eggs. This is where people actually write 2 separate computer codes and merge them together without telling anyone. I remember in the early days (1980s) of the Personal Computer (PC) where the software BIOS (Basic Input/Output Structure on a Read only Memory (ROM) chip) in one computer system included the image of the computer programmer; you could see their face with a few select key strokes. Later, Excel 97 included a flight simulator that could be accessed with various key strokes.
Easter Eggs can be more dangerous if they are included in voting equipment where the tally of the vote can be manipulated. I was on an IEEE committee (P1583) on electronic voting advocating for the independent tally of the vote in the 1990s. We need to guard against the misuse of any computer in every context.
Along with the increased complexity of computer hardware and software are the ways they can be used for criminal purposes. Gathering Personal Information (PI) from your computer about your bank records and accounts, where you work and live, who you associate with, etc. Social media is used for fun by people but they unwittingly give away their PI as they glibly post information they normally wouldn’t share with everyone.
And there is no reason why children should use social media until their late teens. There are predators out there and children have no defense against them other than their parents’ supervision of computers (yes cell phones are hand held computers). Here again, cell phones are called smart phone but they are not smart; you have to be smart when you use them.
AI is not smart. It only has the capability designed into hardware and software that people gave it. That capability can be expanded using SDKs/APIs (Software Development Kits and Application Programmer’s Interfaces). Software companies provide SDKs/APIs to large companies with many users in order to customize the software capability for their particular use. We usually assume the added capability is good but in the hands of a bad actor, it could as easily be bad? Who is stringing the software bits and bytes together behind the scenes?
It is important to recall that People are always behind the use of computers. Not one computer has committed a crime just as no gun has ever committed a murder. The tool used is not the culprit but the person using it (one aspect of the Problem ID); criminals using technology to the disadvantage of others has long been a big problem. We should be concerned about the increased capability of computational machines but not because of the machines but the people who either misuse or abuse them for criminal purposes. The ability to police their misuse/abuse is a growing area of government concern (government contracts anyone?).
Private enterprise has long been aware of the dangers of what we inaccurately call AI and are responding with many hardware and software tools beyond just antivirus programs. ATMs (Automatic Teller Machines) employ a computer architecture once offered by ‘Tandem’ computers. If a computer has 3 independent (tandem) hardware processors, 3 independent software programs on the hardware, and 3 independent sets of data that are identical copies, then the computer is said to be 1 fault tolerant. Given the number of independent processes, if 1 process disagrees with the rest, it is disregarded in favor of the 2 that agree. We want this for our banking data and modern ATMs use dozens of independent processes with a large fault tolerance. This is a method I suggested with use within the IEEE/P1583 committee for the independent tally of the vote.
Apart from computer glitches that added hardware can help correct, how are we going to guard against those who are hacking into your bank account or credit cards? And how are we going to protect our computer software from viruses? There is big money to be made in defending the computing that average people use from the real threats of AI. And that is a big-ticket item. Who is making money off this issue? As some people profit from their criminal activity in computing, others profit from protecting us. Along with all the hype about AI is the bundle of money some will make from its defense. Government funding anyone?
Relying upon computers is what we do because they were made to amplify our limited human capabilities. Without these devices we’re back to using mechanical or electronic calculators (an early form of a computer). So, as we rely upon them and they help us, we are also vulnerable, not only to predators, but the inherent failures of computational systems (hardware and software). Since they are not intelligent, they cannot self-check (some coding tools provide for error checking – not the same). We the People need to have the capability to spot when predictions are non-sensical, when photos or audio files are corrupted, when our banking accounts are compromised, etc.. We should rely upon these tools but not trust AI as something intelligent. The only intelligence is in the people behind the scenes.
*My first earned degree was in computer technology (1983 - AAS Electronics/Computer Tech). I learned how microprocessors worked, learned how to program them, and how to use them for real world applications. I understand computers from the inside out. After getting my BS in Aerospace Engineering at the University of Texas at Austin in 2009, I was accepted into the MS program of Computer Science at Georgia Tech (my mom passed and I later withdrew). I have used a dozen different languages to write programs in school, in my internship at NASA, at my job with data processing, etc. I know computer hardware and software and their limits. And I am familiar with the fact that computers are the dumbest machines on the planet (garbage in – garbage out).
30 years ago, I was a cab driver in Austin, TX and our cab company worked with the tech giants in what was once called Silicon Hills. The merger of activity centered around AI and the ability of the computer system to predict when and where the next cab fare would occur. This project was interesting and I had fund participating with it but learned early on that AI is not what it is cracked up to be. AI then was nothing more than a histogram of past cab activity and it was hit and miss at best.