AI Deep Dive
A look at some of the deeper aspects of AI
I am amazed at the number of people I see jumping onto the AI band wagon. I am seeing hardly any posts or articles against it. I wanted to look deeper. Is AI all that it&';s proclaimed to be? Is it all good?
First a little history. I was in Graduate School in 1980/1981. A professor taught a course titled Artifical Intelligence. Back then, the debate was whether AI could ever truly be intelligent &n-; could it learn? The professor presented both sides quite equally. Then he surprised us (well he surprised me) by stating that computers would never be able to learn the way humans can learn. I had watched enough Star Trek by then to know that he was wrong.
I&';ve been working with computers for more than 40 years. My original idea for my thesis was to devise a program to read flowcharts so that people didn&';t actually have to write code, at least not the simple stuff. That&';s one of the new claims of AI. It will be able to do that. There&';s already template-based websites where you only have to design your website, you don&';t have to program it. This is good for not having to program it. It&';s not so good for websites which might load slowly and use more energy.
I&';m for progress and for making things easier. But let&';s look at all of this more deeply.
On the negative side, just to put it out there: We now have lots of gadgets and much more powerful computers than were available when I was in Graduate School over 40 years ago. These gadgets, etc. are supposed to make our lives easier. Mine at least has become a lot more complicated. I now have to do more myself because companies let me do that. There&';s good and bad in that. But my life isn&';t easier.
I keep wondering if nobody remembers watching Westworld. How about Star Trek Next Generation where there were Data the good android and his evil older brother Lars? AI can turn bad. That might be by accident or it might be due to human interference. Did you read Isaac Asimov&';s Robot trilogy? Those books contain the laws of Robotics which are supposed to prevent robots from harming humans. Yet, something went horribly wrong.
Are these concerns unfounded? Are they trivial enough that we can wait and work on them later?
The promises: AI will make all sorts of things easier. Simplilearn states 18 different things that AI will make easier. There&';s full-self-driving mode in cars which I think is much safer than many human drivers. AI will do trivial tasks for us. I don&';t think the promise is so much that AI will do trivial tasks as it is AI will do onerous, time consuming tasks.
AI will free us up to do other things. Yep, I could get on that bandwagon. It&';s not that I don&';t like working with computers, but there are other things I&';d rather do. And every so often I&';d like to be able to go to the beach.
AI can be our personal assistant. I don&';t want Alexa or Cortana or whoever helping me when I don&';t want their help. I can do it myself! Yet, I can see some advantages in that. On the other hand, it will put human assistants out of work.
AI can write paragraphs for us. It can even write entire stories or do our homework. I use spell check and grammar check a lot. In my current book, I spell the country Turkey the way that country wants us to spell it: Turkiye. Both Microsoft and Google flag that. I also invent words. So, I&';m grateful for spell check and grammar check. But I always double check and ignore lots of suggested corrections.
Taken to one logical conclusion, AI will make it so that we can write in another language without ever having to learn it. That would be good for people visiting another country. But what happens when AI gets it wrong and we end up saying the wrong thing? My text message program often offers corrections which are not what I intended. Often, it offers words I never heard of and I&';ve heard a fair number of them.
Back to the basics: But, what is AI? AI, old definition, is that computers can learn just like humans. The behind the scenes wiring might be different, but the process is basically the same. AI, new definition, computers read and process huge amounts of data very quickly so that it looks like they&';re learning. They might even learn. Most of that learning is pattern recognition which is most of our learning. So, maybe they can and maybe they can&';t. And, there comes a point where it doesn&';t matter. If it looks like learning, it is learning.
But what is it learning? Google remembers what I just queried. But then if I suddenly query something new, it has now way to know that. To be fair, neither does any human. Google keeps thinking that I&';m interested in the old topic and tries to merge the two. Google also thinks that I&';m asking the same question that millions of others are asking. Is that true?
Will AI be bad? Will we use it not so good ways? Most text/chat apps now offer three options as an automated response. Do I want to use one of those or do I want to be unique? Do I want to be like everyone else or do I want to be myself?
I see many posts saying that AI will make it easier for marketers to send me an email that I&';ll be more likely to respond to. I already get too many of those. It&';s good for the marketer, which I am some days. Maybe not so good for the person receiving the email or ad or whatever, which I am every day. I&';m pretty quick at deleting those emails, especially if they look like the last hundred I&';ve received.
Will AI help our brains or atrify them: I&';m all for AI helping our lives. Not everybody needs to be good at spelling or writing. And, we don&';t have to learn everything. But are we in danger of forgetting how to learn? It doesn&';t have to be that way. We can let computers do the ship&';s diagnostics and we just have to double check. We can use our brain to greater extent like those geniuses on Star Trek who can instantly learn Romulan without ever having seen it before. But will we?
Final thoughts: I saw one post stating that AI would make computer narration better. I&';m still waiting for Microsoft Office to offer a better reading voice. Draft2Digital already has a program where certain books can be narrated by computers. I&';m not ready for that. I&';ll still use humans to narrate mine.
When I call a company with a question, I&';m usually routed to a computer who tries to answer my question. The last time, when I said Operator, it said, I&';ll connect you. What happened was my computer then brought up the same chatbot for me to use as what my phone had been using. I eventually got to speak to a human. The computer had no idea what I wanted/needed. It kept telling me I could do that on the website. Yet, the website wouldn&';t let me.
My last point: I haven&';t seen anybody talk about this. How many human jobs is AI going to replace? How will the government support those who lose their jobs to AI. Perhaps AI can solve that problem for us.
Things to think about: Is AI all that it promises to be? What parts of AI should I avoid? What parts should I use? How can it help me? What could possibly go wrong? Are there any safeguards I can put in place?
I want us to us AI to its full advantage. I want us to use it wisely. I want it to be a gift, rather than a burden or a detriment. How do we make that happen?
AI Deep Dive
Add Blog to