ChatGPT and AI - A Cautionary Tale
With all the hype surrounding ChatGPT, and AI in general it is very easy to jump on one of the two main AI bandwagons.
?There are the “Doomsayers” who think that AI is going to wreak all kinds of havoc in the art world, academia, replace all of our jobs and eventually become sentient and enslave humanity (you youngsters should watch 2001: A Space Odyssey for a preview).
?Then there are the “Worshipers” ?who kneel at the temple of artificial intelligence believing it will solve all mankind’s problems.
?For most of us, who don’t live on the cutting edge of technology, AI has been a vague concept that few of us have consciously experienced.?We have used AI unknowingly in all sorts of arenas from online shopping suggestions to airline ticket pricing, and have contributed to the AI knowledge base through our online searches, shopping habits and social interactions.
?ChatGPT has moved AI from unconscious usage to conscious usage.?It is being referred to as the “Google Killer” with it’s brilliant responses to our questions posed in natural language.
?Like many others, I took ChatGPT for a test drive and was truly amazed at the results.?Over the holidays I asked ChatGPT to write a Hallmark movie about Christmas in Asheville North Carolina and the resulting story was spot on in it’s rendition of Hallmark tropes and Asheville culture.?I was sold!
领英推荐
?But how about using AI in real world situations??Is it reliable??How do we know?
?I am a home automation enthusiast, and use highly customizable open source software called “Home Assistant” to run the automations in my home.?I subscribe to a Home Assistant Reddit forum where people frequently ask for help setting up all sorts of automations.?It’s been a fantastic resource and has saved me, and many others, hours in trying to solve problems by getting help from experts or others who have solved similar problems.
?What does this have to do with AI or ChatGPT??Well, recently people have taken questions asking for help on the forum, running them through ChatGPT and postings the responses as answers to the questions. Some of these responses have caused ?systems to crash, or have all sorts of other unintended consequences.
?The moderators of the forum have now banned people from posting ChatGPT generated responses on the forum.
?So, what does this mean for ChatGPT and AI for the masses? It can be extremely valuable if the user has the requisite knowledge to understand when the response is incorrect.?However, if one takes the results as gospel, without a solid understanding of the results, it can quickly transform one from a Worshiper to a Doomsayer.
Visual Artist, Photographic Educator and Arts Community Leader
1 年I agree that you shouldn’t jump on the bandwagon, but you should also check it out and just not write it off as well. I’ve tried it regarding writing an artist statement. I found that the user has to have a really good grasp on what they’re trying to say. I tried different ways to get it to say something different, and it always gave me the same result in the end. I had to edit it with information that was more in line with my thesis. Also, another colleague of mine, tried it with medical information and found that it did not have access to certain databases.
SIG, SIG University - leader of the world's largest global association for sourcing, procurement and risk professionals; leading through education, networking and influencing.
1 年Great insights. Don’t jump just because it is new and shiny!