MT 43 News Blocks Example Article

Example Article for Tech Talk

Yet Another AI Article

Author: Victor Sample Vic Sample: MT43 News Treasurer

Victor Sample

AI is one of the hottest topics these days. I continue to get a large number of AI articles in the newsfeed on my phone; of course, since I read them, they keep showing up in my feed – a nice feedback loop!

As I have mentioned before, I am not against AI – I use AI almost every day. However, I am careful about the results. I always try to verify any information I get from AI. In my prompts, I am always careful to state that I do not want any “hallucinated” information. Of course, that doesn’t keep AI from hallucinating, but I have found it helps cut down the number of hallucinated answers.

Recently, I have read articles about 2 different law cases that were thrown out of court, and the lawyers were reprimanded for submitting briefs that were generated by AI and were full of errors, irrelevant citations, and even made-up legal precedents. In one case, the attorney submitting the brief was reprimanded and FINED. In the other case, both sides submitted AI-generated briefs; the case was thrown out, and both law firms were sanctioned by the judge.

I have found AI to be very useful when I ask a technical reference question; i.e., a question that I could find in a technical reference book or in a technical document. For instance, I was using the AI tech support a question about options for updating a crontab table without using the standard Linux text editor (which I detest). The AI agent correctly gave me a step-by-step list of what I needed to do to add a crontab entry using a CPANEL interface that is much better than the standard Linux text editor. Clearly an AI success story.

However, when I asked the best way to do an intricate process of creating a long-running Linux task without “hanging” my webpage, AI was not very successful. We went down a lot of dead-end, very complicated paths before I changed my approach and found an easily implemented answer to the question.

Again, AI does not represent intelligence the way humans exhibit intelligence. AI formulates an answer by mathematically deriving the most PROBABLE answer. In the case of the law firms, the mathematically derived “most probable” answer led AI to generate erroneous answers and even create citations and precedents that did not actually exist. AI does not lie – it’s just that many times the mathematically most probable answer is not the best answer.