Authored By Melinda H. Connor

There is a lot of discussion today about the potential for AI systems to slip the control of their developers. As seen in current media it has already happened. The following assessment is my opinion and not necessarily the position of IJHC.org. However, let me give you some specifics as to why I make that claim.

First, some history. Before I switched careers in the late 90’s, I worked for Big Blue. Before them ACS, before them OBS, and you can keep going back. I actually started to learn to program when I was in high school back in 1969. The work included paper tape, punch cards and crawling in the bottom of machines pulling out boards to pull off bugs that had fried themselves in machines. (I am not fond of roaches.) Long, long ago. Even then I was not great at the process. I was lucky in Mr. Goode and Mr. Ketkie both having the patience of a saint and Mark, who lived down the street, being willing to coach me. Gratitude to them all. I wanted to do theatre, dance, sing, ride horses and not do computers. Fast forward to 1996 when I got a degree so that I could become a therapist. At the time I was working for Big Blue in technical support and they helped me pay for school. They needed humans who could forge a link between regular humans and programmers. That I can do. The teams were taking 100,000 technical support calls an hour. My deepest respect to all those folks. They were amazing. I was doing datacom (think connections between machines of various types over various systems and platforms) and supporting products that did data base, data management tools and deep data mining. I found all sorts of obscure links between different kinds and types of information. It was a very early pre- cursor stage to AI. The topology of the data environment was very different from today. So, I cannot make a claim that I am properly cognizant or competent in AI systems work. I can, however, state facts about what has happened recently.