Artificial ignorance

Recently, I pondered if I ever would understand artificial intelligence (AI). Like many people, my understanding of AI didn't go much further than the virtual assistant apps like OpenAI's ChatGPT, Apple's Siri, Amazon's Alexa, the Google, Yahoo and Microsoft Bing search engines, and the self-driving Tesla car. I decided to do a bit of research to learn a bit more about it. Little did I know that I was to go down the rabbit hole of computer science and engineering once again.

My mistake was beginning my search with the jargon of research journal publications. However, "Introduction to Artificial Intelligence" by Abdulaziz A. Almulhim of Saudi Aramco in the International Journal of Innovative Science and Research Technology did get me off to a good start. Almulhim presents an interestingAI timeline, beginning in the 1950s and going to the 2010s. Definitions and AI success stories have come a long way since then.

A major takeaway from Almulhim and other references is the three different types of AI, based on their capabilities: First, Artificial Narrow Intelligence (ANI or "weak AI") which is everything we have now capable of solving one single problem or executing a single task really well. Second is Artificial General Intelligence (AGI or "strong AI") which is a hypothetical system we don't have now, but the subject of dozens of research and development projects worldwide. AGI is reminiscent of the android character Sonny NS-5 in the 2004 movie "I, Robot," adapted from Isaac Asimov's book of the same title. The robot had strength, speed, a flexible face and programming that granted him free will. The third kind is Artificial Super Intelligence which will be able to surpass all human capabilities, which in my belief we shall not have until we leave this present reality and join the God of the universe.

There is a lot of hype about AI tools. There is no standalone AI. It is always software running on hardware and interacting through an interface. There are always humans inputting massive data sets to get it going. More importantly, we would be at a loss if the Internet goes down because of weather interference or not having an electrical power source or not having enough ground-based antennas or satellites. I have read that the space above us is getting crowded with satellite clusters, and crashing satellites are inevitable, as the five- to 10-year life spans of the thousands of satellites ends and they become space debris, but that's another story.

There are several free introductory AI courses online, each with differing time commitments. The Elements of AI from the University of Helsinki is nice, as it is self-paced, focuses on AI basics and offers a free certificate upon completion. For myself, I have ordered "Intelligent Ignorance: The Promise and Peril of AI" by Miles Kenner. After reading, I will generate another article. In the meantime, I'll head over to Stanford University to check out a cognitive shocker I read about, in the form of a claim that a Stanford AI model can predict a patient's death with up to 90 percent accuracy, and that this will improve end-of-life care planning. I'm an empirical doubting Thomas. Seeing is believing.

The author (wrjones@vsu.edu) published the novella "Beyond Harvard" and teaches English as a second language.