152 AI what is AI - editorialpinups

I have been involved in the discussion of AI since my high school days in the early 1970’s.  Process automation was one of the main uses.  A second use has been embedded in data retention and access. The early stages were involved in the robotics in manufacturing and certain consumer services.  One of the first that I saw was the car wash. There were different car wash technologies.

The first automation was to create a long conveyor which crabbed the car and dragged it past the various functions that at one time would be done by a person.  Spray the water, then spray the soap on the car, next brushes spinning on the sides and over the top.  The car gets sprayed with water again then it gets ejected from the conveyor turning the car to a person.  The decision tree was mostly dedicated yes no and here or there.  There were various persons watching the process and if there were problem to occur there was the emergency shutoff switch.

The next type of automatic car wash had the same functions for applying soap and water, running brushes, and at the end optionally blowing the water off the car.  But the big difference was as the car driver you drove through the tunnel with the tires in a rut system to keep the direction, where this was set up and drove according to the speed of moving lights on the side wall.  For the operator this was a less expensive operation as the conveyor was more expensive to install and maintain. The light system had a smaller physical footprint.  If there were problems there were manual emergency shut off switches along with sensors and light beams that could turn off the system if the light broke.

This became college discussion for me in the late 1970’s. The beginning of computer aided learning started kicking in the mid to late 1970’s. Computer programs were written to do large comparisons, then making decisions.  Computerized test grading came along with the beginning of computer games including chess, caverns and tic-tat-toe. Development of sensing devices that could pass data from machines and items of data gathering started growing.

My involvement in computers and Data Processing involved fund raising solicitations, then on to accounting, inventory systems, and office automation.  The goal was to eliminate repetitive tasks in the office.  There was much duplication industry wide during that time. Companies created computers and programming languages, then businesses defined their needs and wrote their own systems.

Along the way, many people who were doing this programming got entrepreneurial noticing the consistent parts of accounting systems and the likes of Great Plains, Open System Inc, Peachtree, then eventually Intuit.  They made detailed reporting and data capture functions.  There were no spreadsheets, word processing was starting on specialized machines, data communications over telephone lines was growing.

In the early 1970’s there was data over telephone lines.  It took expensive dedicated equipment.  The teletype was one of the early devices.  I had the opportunity to do my first computer programming using the teletype.  I made my paper tape containing my computer program, dialed the remote computer which processed the program and shortly the computer would call back and have the results printed.

Next came 1981 I worked for an organization that had remote fertilizer plants in small rural farm towns  which used a centralized billing system.  A process was created to dial the fertilizer plant computers from the central location then having the two computers talk and pass the data to the billing system.  A few years later I had the opportunity to write a process on a different platform transmitting heart data for pacemaker analysis.

With the advent of home and gaming computers a company that I had a contract for accounting system work told me about a process they had where they needed to monitor a pressure valve.  When it hit a condition the valve was to be adjusted.  They bought a bunch of $100 Commodore 64’s when small business systems were going for thousands of dollars.

All bits and pieces of AI.  This was before 1990.

While on a contract at 3M in the late 1980’s I had the chance to lunch with some people developing AI.  They said it was the beginning and they were programming the teaching of very small items.

I had been doing amateur photography since 1988 after my mom gave me her single lens reflex camera when she went to simpler point and shoot system.  I had made an arrangement with a company in a previous city I had lived that processed film to make my prints, but also scanned negatives and printed negatives from digital files.  They had told me the intermediate part of the process was using Photoshop.

Fast forward for me to 1996.  I started to do professional photography using film. Digital cameras had not made their commercial presence. If you wanted to touch up a photo, you either had to paint the printed picture or yes, paint the negative. But remembering that vendor from my amateur days I could send them the negative and tell them want I wanted removed, have them make the negative and my print then sending me both.

Eventually I learned Photoshop and its free cousin GIMP.  I learned how to smooth skin, get rid of pimples and blemishes, red eye and glare. One day I see this program called Portrait Professional.  It had this programming to look at the person in a picture, have you tell if it as a woman, girl, man or boy, mark the area that you want smoothed then go, it will do its thing.  You get a changed picture.  You can then tweak it, in most cases make it look a bit by the original person and you are good.

Another program that I started to use is Topaz Remask.  The purpose is to mark around a subject on a picture and cut the background off.  Mark, computer, tweak, save. A nice tool.

Bur recently, looking at the current web sites, Portrait Professional uses AI.  Gimp has AI functions.  Topaz Remask AI.  My question, what were they doing before?  It all looks pretty much the same.

I had worked for a company called IP Soft.  They went from being an IT help desk and a support company to a company that supplies AI functions for Help Desks in IT, HR, Medical and a number of other industries. Founder and CEO Chetan Dube once said that to give yourself a raise, automate your tasks which allows you to do other things.  A good thought, but…..

Another aspect of AI concepts were an experience that occurred when I decided to get a cup of coffee from Casey’s one day.  Casey’s is the gas and food store that has good inexpensive coffee.  Traditionally coffee at Casey’s was brewed in larger urns through the drip process.  Usually 30 plus cups of coffee were made at one time then it was hoped to sell them in specified time period and would then dispose of the excess as the taste would not be good.

This store had converted to the glorified Kuerig system.  There are three canisters with un-ground coffee beans behind glass on the top of the unique machine.  The three canisters are labeled with the type of beans in the system.  You tap the control glass and choose which canister you want the beans selected.  Next select the size of the cup.  Choose whether to leave room for cream and sugar or fill completely.  Hit start and watch.

First you will realize that there is faith that the beans selected is used. Visually you cannot tell.  Second, it is by faith that the water will stop when the conditions are met.  It is faith that it will choose the right coffee beans, that it will know how much to fill the cup.  But it is reality that there is the need for the emergency shut off button.  Using the wrong beans is not something the decision makers cared about.  Cleaning up excess coffee on the counter and the floor was a desire to avoid.

Understand that these are a few examples of AI in the past and current.  There are many more and will be even more uses of AI and process control. AI will be used to sifting through large amounts of data to give informational results. Interpreting these results of AI must be done with either extreme faith or intense scrutiny.

Intense faith is an easy choice if the source of the AI has been given good reason to be trusted.  For many, that leaves out Google and Microsoft.  However, best practices state that it is still too early in the game to not scrutinize results.

AI will be able to give results that people will not like.  We see that thought process keeping in as legislators are talking about possible legislation to try to control the results of AI.  They fear the results will expose accepted dialogs as falsehoods.  

Understand that as AI is only as good as the programming resources and components used to create.  AI being a learned process must have a proper basis of process to start, then must continually learn in a valid manner.  This process must be evaluated just like students going to a school and learning.  Beware that the results my just not be what is wanted or expected.  Governments and political parties will try to legislate even ban AI tasks and functions that will not allow them to push narratives.

AI must be allowed the same protections that the First Amendment allows for speech. But remember how the First Amendment is incessantly challenged.  There will be groups who will say that the protection does not apply in certain cases.  There will be court challenges. Judges who haven’t a clue will make decisions or jury rules. Lawyers will charge ridiculous fees to represent the case.

There is the question of intellectual properties.  The risk of AI is that in its loosest of definition and evolutionary capabilities could learn, mimic, and turn itself into another AI product. Good luck in determining who owns what and who created it.

Welcome to the world of AI!   Again, exactly what is AI?