Explanation of where I got my name.
-
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
Every year I have to take some compliance training for work. It rarely changes much from time to time. Last year it was changed to AI. I find the voice quite annoying, with it's subtle mispronunciations and such. The thing is, it's just reading what someone typed in the computer. Anyone could voice it better than the computer, in my opinion. It made me wonder if there is a data storage advantage or something between Joe Blow reading into a mic and letting the computer do it. I'll say, every time I complete training I praise those still doing it the old way and send a nastygram to those using AI.
Sounds like you were paying more attention to the speech-synthesized version than if it had been recorded by a human.
Bing search does that now. And it provides the documentation for the sources. So I don't see where AI is any worse. It all depends on how it it programmed.
As for wrong information, we have already entered an age where "wrong information" depends on the eye of the beholder. This is why Bing's providing documentation in the form of footnotes is an example of a positive way of using AI.
Again, it's about duplication. We're inclined to believe things if we can find multiple sources that say the same thing. What does AI do? It scrapes the internet for information. Unless the information online is all correct (and we know that's not the case), this is garbage in, garbage out, but times your ZIP code.
Nonsense. Facts are facts. The whole "My ignorance is just as good as your knowledge" attitude---called out by Isaac Asimov 44 years ago---is an aberration of our society, not an evolution.
If 13 articles all say the same, wrong, thing, then the only thing that's helpful is human beings fact-checking that (which gets harder as AI pollutes more of the knowledge base), and sharing the findings with the folks working on AI to make corrections and hopefully create a more reliable product.
Especially when the LLM is scraping from sources like Wikipedia.
I'm ignoring him and can't see what he wrote, but Boombox frequently likes to defend feelings over facts. I'm willing to bet that's continuing here.
The interesting twist to AI learning has become copyright violations and plagiarism. Reddit suing Open AI for Chat GPT mining their site for information and opinion is a problem. Same with Microsoft's AI search engine; Bard.
Ideally, if you use Wikipedia, a source has been provided, and if you're lucky, the source is accessible online.
However, the information could be more often than not unreliable or deficient because it's being edited by individuals with no actual knowledge of the topic.
The experts don't tend to edit, and when they do, they sometimes use their own knowledge thinking that's enough, not realizing we don't know who they are or see something that has been verified.
Then that's just another reason why Wikipedia is a wholly unreliable source for any information online.
No, that's unfair. Wikipedia can provide very reliable information. It can also provide garbage, or opinion disguised as fact. It all depends on who's contributing, their level of subject-matter expertise in the particular topic, and who's doing content moderation. It's the inconsistency that causes the problem. I can only know for sure if an article is accurate if I'm already a subject-matter expert myself, and if I am, why do I need Wikipedia?
So if I'm understanding you correctly, the best information will be available in New England (with Zip Codes 0xxxx), and the worst on the West Coast (9xxxx). And God help the people of Ketchikan.
I don’t think that’s unfair at all.
If you know someone who tells you both the truth and total crap at different times on different subjects, would you consider that person “reliable”?
Kelly said Wikipedia is a wholly unreliable source for any information online. (Emphasis mine.) It's a crowdsourced site, and the reliability is dependent on who's contributing, and *what* they're contributing. I've found thoroughly useful information on scientific, medical, anatomical, engineering, etc. topics, contributed by bona fide subject matter experts (SME's). State-of-the-art information that can easily be fact-checked and corrected, if necessary, by other SME's. Where we get into trouble is when people get out in front of their shoes, providing opinion in the guise of facts. Or worse, someone who doesn't know what they're talking about weighs in as if.
In the case of any individual person, it depends on where they're coming from. IF they're a SME on the topic they're writing about, and their writing indicates they have enough facility with the written English language to write precisely, then yes, I trust them. If they're being a blowhard about [e.g.] politics, then it doesn't matter how much of an SME they are about the Tao Lepton particle, or the structure of a brain cell, or digital signal processing, or the Constitution. (I chose those subjects intentionally, since I know or knew people who are or were experts in those fields. In the first case, that person had the Nobel in Physics in that topic. But opinions are like armpits, everyone has at least two, and nobody is immune from bias, prejudice or plain old bullshitting.)
If you use Wikipedia unquestioningly, it's no better than using ChatGPT unquestioningly. Or for that matter, the Encyclopedia Britannica.
(BTW, I presume you've left the building already, as today is 31 January? Enjoy your freedom.)
Nope. Today’s the last. Thank you!
For those of us who don't follow every thread, what are you talking about, please?
But that would make the reliability amount to a roll of the dice, wouldn't you agree? The old saying; 'Even a broken clock is right twice a day' comes to mind. If you know the clock is broken, would you depend on it as an accurate way of timekeeping?
The fact we have people here on this very site who are contributing and editing content with no other knowledge of a subject except what they look up online, speaks volumes about how unreliable Wikipedia is. Many kids of K-12 up to college-level adults have been burned by relying on information listed via Wikipedia.
Again, dice rolling that information contained is correct? How would anyone with the common sense God gave a goose rely on something like that?
That's easy, you wouldn't. But is Wikipedia looking for donations to support something that is intentionally an information portal that's a crap shoot away from being correct information? If so, then it hasn't been doing the Internet much good all this time, now has it?
Sorry, Semoochie:
Thread 'An update to my introduction'
https://www.radiodiscussions.com/threads/an-update-to-my-introduction.769125/
What I said about "wrong information" being dependent on the eye of the beholder -- whether it's an aberration of society or not, is accurate. During the pandemic in 2020, if I would have posted what the Biden Administration concluded -- less than a year later -- about the source of the virus, including using their same exact statements, I would have been blocked on social media, thanks to their expert "fact" checkers. Many who were posting the same exact statements that the Biden Administration later duplicated in their official statements were blocked for disseminating "misinformation" in 2020. Blocked by algorithm, basically.
The facts, or information didn't change at all. But it was the way it was perceived. The same exact information that was deemed unacceptable in 2020 was suddenly deemed acceptable later in 2021.
AI could be used in much the same way, of course. Limit what the internet user sees. We already have a bit of that today, AI, or no AI. I think it's a valid concern, but I don't see AI itself being the problem. It depends on how it's programmed -- just like today's search engine results -- which most people use in research -- all depend on how the algorithm is programmed.
AI does what anyone does: scrape the internet for information, and the way it does it depends on how it is programmed. As for 'duplication', even some news sites just quote other news sites when writing their stories -- which is basically the same thing. When you have a country of 340 million and only ~40K journalists, you're going to get a LOT of duplication and lack of adequate research on stories. AI could indeed make that worse.
It's always up to the individual to check out the sources in the documentation, something most people simply do not do. Instead, they rely on their own confirmation bias. It's human nature. We see plenty of that here on RD, and on social media as well. Even in radio (conservative talk radio being a classic example).
The fact is that most Americans aren't taught to search out documentation, how to do research, or how to use critical thinking. AI isn't going to change that.
Which is why AI can go either way. If it's programmed to seek out diverse documentation, the results may be more accurate. If it's merely programmed to bring up the "top" results from a search engine (which depend on algorithms themselves), the results will probably be skewed.
Ultimately, it's up to the user to use their critical thinking skills.
You mentioned that AI could "pollute the knowledge base". We've already got a lot of that with the major search engines -- sites that probably 90+% of people use to research news and other information. None of the sites seem to bring up the diverse results on the first 3 or 4 pages like they brought up in the 2000's, before the companies decided to monetize the search engines more effectively. There are major items, surveys, and other data that I recall bringing up on the first or second page on one of the biggest search engines in 2005 that I can't even find today, even after searching through tens, if not hundreds, of pages of results. A lot of the time it's page after page after page of the same repetitive crap, rehashed. It's like there has been a limit placed on what search engines will provide in results. Maybe experts in Boolean search terms can get better results -- most of us aren't experts in Boolean search term use. I know I'm not.
So, if search engine algorithms already give most people the same regurgitated pablum as results, how is AI going to make it any worse? After all, AI is just another algorithm, just another computer program -- albeit a very complex one. It's just a tool. It's how the tool is used that matters.
So, I get your concerns, but I think in a way we're already there.
- Status