That's actually part of the potential concern: AI applications that build AI applications. For example; I can task ChatGPT4 to write an application with a lot of detailed complexity in Python, and it will do it in a matter of seconds, verses an engineer in days.
-
Get involved.
We want your input!
Apply for Membership and join the conversations about everything related to broadcasting.
After we receive your registration, a moderator will review it. After your registration is approved, you will be permitted to post.
If you use a disposable or false email address, your registration will be rejected.
After your membership is approved, please take a minute to tell us a little bit about yourself.
https://www.radiodiscussions.com/forums/introduce-yourself.1088/
Thanks in advance and have fun!
RadioDiscussions Administrators
Artificial Intelligence Replaces Radio Announcers
- Thread starter davideduardo
- Start date
As more and more "humans" lose their livelihood to AI technology, there is going to be a growing backlash and resentment towards AI from the public. I predict this backlash will reach a point that government will start to legislate AI to gain control over an eventual takeover of society.
As for radio, in markets where there are 2 or more format competitors, and one station using AI on air to some extent, the other could vow to not to use the technology, and could promote that fact with listeners, especially if there is growing resentment from the public.
I can just hear those liners now..."No AI here, all human, all the time! 91X, AI free since 1983!".
As for radio, in markets where there are 2 or more format competitors, and one station using AI on air to some extent, the other could vow to not to use the technology, and could promote that fact with listeners, especially if there is growing resentment from the public.
I can just hear those liners now..."No AI here, all human, all the time! 91X, AI free since 1983!".
The essential failing of AI is that it will, along with correct information, find and use incorrect information on the web.
For those of you who wonder why I'm such a (insert epithet here) about getting the facts right, it's this. I used to think of it as "someday, someone will read the wrong stuff and believe it". We've blown past that. Now it's being assimilated into material through AI.
For those of you who wonder why I'm such a (insert epithet here) about getting the facts right, it's this. I used to think of it as "someday, someone will read the wrong stuff and believe it". We've blown past that. Now it's being assimilated into material through AI.
YES!
I've been testing Bard. It always presents answers that are cogent, clearly written, compassionate, logical ... but are often 100% incorrect.
I applaud Google for warning that Bard is in the experimental phase and that its information may not be accurate.
Will users heed that warning? That's the question.
With few exceptions, modern AI 'large language models'' are designed to respond with 'tokens', or a potentially reasonable-sounding response in accordance with what it believes the questioner wants to hear/believe:
Great paper on the subject from Cornell: [2304.00416] Towards Healthy AI: Large Language Models Need Therapists Too
Should’ve said it was a British computer
I have a Richie Rich comic book from the 60s where the robots learned to do everything themselves and then started trying to take over.
Yesterday I was listening to the radio and the DJ, who was really there, said he was pushing all these buttons and wasn't sure what was happening. He saw green lights and red lights and didn't know until the song started playing whether he had done everything right.
Yesterday I was listening to the radio and the DJ, who was really there, said he was pushing all these buttons and wasn't sure what was happening. He saw green lights and red lights and didn't know until the song started playing whether he had done everything right.
A computer could do that too if you gave it the right instructions.
You mean like ENCO or RCS?
The smallest and easiest part of the app you described is the coding. A monkey with a banana can learn to code. The human that figures out what is needed, how it will function, how it will protect itself from all the malarkey out there, whether it will be cost effective and deal over time with supporting and adjacent apps and hardware is the real test. (And there is a ton more considerations I won't go in to here.)
Back in the early 70's there was a product that used decision tables to generate COBOL programs. One of its biggest shortfalls was the use of double negatives to make decisions. Very few people on Earth can accurately follow a double negative path and, needless to say, it did not last very long at all (at least in the USA). What was its big selling point? Why, cost avoidance of course. No need for those pesky and expensive programmers.
AI smells like this shallow thinking all over again and I predict a similar ending. Of course, if you financed the cost of software development with your crypto money you'd be in Fat City, right?
I hate when the monkey on the early shift leaves those banana peels on the keyboard...
WZQR-LP 100.7, a low-power FM in Bokeelia, Fla. (next to Fort Myers), has a standards/big band format. The owner spent a lot of time in Chicago and evidently fancies the WGN of the 1960s-70s, as one of his announcers uses the name Eddie Hubbard. The station also has an all-night show with old clips of Franklyn MacCormack, including the intro to his old "Meister Brau Showcase."
In recent days, WZQR has brought the late Mr. MacCormack back to life through AI. Just a few short segues in what is remarkably close to his voice. I would think if there was still a MacCormack estate, it would have a case.
In recent days, WZQR has brought the late Mr. MacCormack back to life through AI. Just a few short segues in what is remarkably close to his voice. I would think if there was still a MacCormack estate, it would have a case.
Last edited:
Third world problems.
Banana republic?
People read the wrong stuff now and people believe it. How will AI be any better or worse.
It'll be in more places, essentially duplicated and thus giving itself the appearance of information from multiple sources.
And, as reliable sources of information give in to the temptation to be current or the economic pressures and accept the technology, there's a greater chance of the wrong information being given credibility even among smart readers.
And sour ciggies in the ashtray.
- Status