The author is chief engineer at Crawford Broadcasting Company-Birmingham.
There’s no doubt in my mind that if you are an engineer type you have not only at least heard of Artificial Intelligence (AI), but you’ve also seen some of the biggest science fiction blockbuster movies that pit “man against machine” and the intelligence behind the machine. The truth is that we’ve heard about it, but many of us have never had an opportunity to really work with it, until now.
A couple computer people that I follow on YouTube made several videos about ChatGPT that inspired me to test it out and see where the technology is in its current state.
Essentially, ChatGPT is the Google search engine on steroids. While a typical search of the internet can provide you with a vast number of sources for information, it is up to you to pull the information together in order to meet your requirements. You can ask ChatGPT to do any number of things within a given set of parameters that you provide, and then watch as it churns out the product.
For instance, chief engineers in our company write articles for the Crawford Broadcasting Company Local Oscillator publication that share our information about problems and solutions. We had an antenna lightning strike in December and I asked ChatGPT to write me a 500 word article about the lightning striking the antenna, our use of a drone to see what antenna bay was hit and what our tower crew did to fix the issue. Thirty seconds later, I had 500 words and a fairly well-written, albeit soulless document that I could have used.
The really curious among you are probably asking yourself, who says he didn’t use ChatGPT to write this article? Spoiler alert, you would have been able to tell that it wasn’t my normally witty, astute writing style…or could you? At this point, your decision is probably hinging on whether you think I even know what the word “astute” means.
The power of ChatGPT is in how fast it turns out a request. What if I didn’t like the way it was done? ChatGPT gives you the opportunity to regenerate the request, and you’ll get another well-worded document. I did, in fact, show our creative writers what it can do. We have a local auto shop that advertises with us. I told ChatGPT to write a 60 second radio spot that included the name of the advertiser and their location. Again, in 15 to 20 seconds, our creative writers agreed that they could have it voiced and put it on the air with minimal adjustment to the content of the ad for the shop written by the AI.
[Read More Guest Commentaries Here]
I’ve read a number of stories about students using it to churn out five paragraph essays or research papers written in MLA formatting. I’ve also seen a demonstration of the AI engine creating pictures by simply giving it a description of what you’d like to see. I don’t write scripts or code, but I’ve needed a Windows powershell script and a Python script to do some testing on an issue we’re having, and ChatGPT cranked out both for me. I handed them to a couple people I know that are programmers and, with only a few minor tweaks, adding information to some generic variables in the code, they said the scripts were solidly written.
So what is really happening in ChatGPT? In short, their AI engine has been trained with human responses and then “taught” with positive and negative reinforcement how to respond to various questions. It integrates all of the search data and all of its training into a comprehensive answer to queries it is given. The intensive search and computing resources necessary to complete each task require more horsepower than a standard search. The added data crunching costs more per job to accomplish what you ask it to find or do. In fact, ChatGPT has gotten so popular that their current computing capability is having trouble keeping up with the sheer volume of requests that it it receiving from users.
You may be reading this and wondering whether this will be another computer fad or if it is just another novelty that will pass like AOL’s “You’ve Got Mail” email notification. On the contrary.
Not only has Sundar Pichai, CEO of Google, told their teams that they were in “code red” to roll out products that incorporate AI into their tools, but Microsoft (an investor in OpenAI, ChatGPT’s owner) also plans to feature tools in their Bing search engine by the end of March. So, “Big Tech” is about to jump in the ring to bring the technology to everyone and probably lower the cost per search for the service as well.
In case you’re worried about the implications that ChatGPT presents, I did ask ChatGPT if it will eventually take over the role of the radio engineer and this is the response I got:
“It is possible that artificial intelligence and machine learning techniques could be used to assist with tasks performed by radio engineers, but it is unlikely that they will completely replace the need for human expertise in the field. Radio engineering involves a complex mix of technical knowledge, problem-solving skills, and practical experience, which cannot be fully replicated by a machine. In addition, there are many aspects of radio engineering that involve decision-making and creativity, which are difficult to automate. While ChatGPT and other artificial intelligence systems may be able to perform some tasks related to radio engineering, they are unlikely to fully replace the need for human radio engineers.”
Of course, if you’ve seen the movies, isn’t that how all of the AI stories unfold?
What are the implications of this new tech for radio and radio engineers? Comment on this or any article. Email radioworld@futurenet.com.