Chat GPT - the good and not so good

coffeeaddicted · Dec 14, 2025

Hazel123 said:
So I quite like Chat GPT sometimes - it's good on legal matters, the odd bit of coding and quick in depth answers. Tonight I discovered it is not very good on the topic of Christmas cakes ............. When querying how to cool one overnight and whether to cover it or not, it said "I often just put a tea towel over it and leave it out overnight".

Clearly Chat GPT has not done that! So I asked it - so you've actually made a Christmas cake and done that? Caught it out - er no sorry I was just quoting that from this web page.

Since chatbots have only a 50% accuracy, i doubt i would trust what it says.
I used ChatGPT but switched to Le Chat which i use sometimes but not for really essential questions.

Infographic: How Accurate Are AI Chatbots?

This chart shows the share of answers from the following AI assistants (free versions) that contain significant/minor inacurracy (in Dec. '24 vs. May-Jun. '25)

www.statista.com

gunrunnerjohn · Dec 14, 2025

One significant issue that everyone can see is it's impossible to trust anything you see on social media or places like Youtube. There's so much AI BS content that it's totally swamped any useful content of those platforms.

Scannerman · Dec 14, 2025

AI has taken disinformation to a whole new level. Not only is it a game changer — it's a world changer. I cannot bring myself to believe the change is entirely for the better. A propagandist could not ask for a better friend. Thanks to AI we can no longer trust anything we see, anything we read, anything we hear. AI can be made to state anything anyone wants and, thanks to AI, will be passed off as factual information. Chat GPT is currently the grand-daddy of nursery crimes. Future generations will learn to live and love a lie like never before. The more they lean on AI to do their thinking for them while they continue to ingest and regurgitate fictions and half-truths generated by these programs the more ignorant and mindless they will become. Meanwhile, those students who take it upon themselves to roll up their sleeves and think for themselves, innovate for themselves, do their own white paper studies and fact checking etc. will be left in the dust because obtaining and processing data in this fashion is pain staking and time consuming. They will not be able to compete with the machine that all the world so readily embraces. Besides, (it will be argued) people make mistakes too.

FrancoDT · Dec 14, 2025

Scannerman said:
AI has taken disinformation to a whole new level. Not only is it a game changer — it's a world changer. I cannot bring myself to believe the change is entirely for the better. A propagandist could not ask for a better friend. Thanks to AI we can no longer trust anything we see, anything we read, anything we hear. AI can be made to state anything anyone wants and, thanks to AI, will be passed off as factual information. Chat GPT is currently the grand-daddy of nursery crimes. Future generations will learn to live and love a lie like never before. The more they lean on AI to do their thinking for them while they continue to ingest and regurgitate fictions and half-truths generated by these programs the more ignorant and mindless they will become. Meanwhile, those students who take it upon themselves to roll up their sleeves and think for themselves, innovate for themselves, do their own white paper studies and fact checking etc. will be left in the dust because obtaining and processing data in this fashion is pain staking and time consuming. They will not be able to compete with the machine that all the world so readily embraces. Besides, (it will be argued) people make mistakes too.

I've said it before and continue to say AI is only as good as the sources that train it and feed it information. AI models will mirror and often amplify the biases present in their source information. AI will discriminate, spread misinformation, and bigotry because that is what it has been taught by inherent human bias throughout our history.

coffeeaddicted · Dec 14, 2025

Something happened. Not sure if that happened to someone else.
I am rather confused.
I was watching yesterday a podcast ( i thought it is one) that was from Michio Kaku about why we never can go to mars.
Now, today i wanted to watch another episode but it's gone. My history has no knowledge that i ever watched it.
I noticed a lot of channels that have Mitchio Kaku but seem rather strange. Perhaps AI generated. I have the feeling that what i was listening wasn't Michio Kaku but rather an AI that sounded like him.
This is so bad. Most channels i seen are most new. 1 day, 2 days. maybe a month. Not older.
I feel i am loosing trust in YouTube.
Anyway, it actually scared me.

Hazel123 · Dec 14, 2025

coffeeaddicted said:
Something happened. Not sure if that happened to someone else.
I am rather confused.
I was watching yesterday a podcast ( i thought it is one) that was from Michio Kaku about why we never can go to mars.
Now, today i wanted to watch another episode but it's gone. My history has no knowledge that i ever watched it.
I noticed a lot of channels that have Mitchio Kaku but seem rather strange. Perhaps AI generated. I have the feeling that what i was listening wasn't Michio Kaku but rather an AI that sounded like him.
This is so bad. Most channels i seen are most new. 1 day, 2 days. maybe a month. Not older.
I feel i am loosing trust in YouTube.
Anyway, it actually scared me.

Maybe he just deleted it. That's happened before - or it didn't get saved in your history. I don't think AI creates false podcasts of real people

TraderGary · Dec 14, 2025

I tap my Copilot key many times every day, but it's always with technical questions. I've been using Copilot for technical help and explanations for the past year and a half. Copilot footnotes its sources, and I've often clicked on the footnoted information source and I'm confident I'm getting good technical information. Copilot has helped me immensely in my C# programming with suggestions and explanations. For how I use Copilot, I'm quite satisfied.

coffeeaddicted · Dec 14, 2025

Well, i found another channel and its not him.
I think that the channel i watched was not him speaking but an AI from a transcript perhaps.
He doesn't have a YouTube Channel.

Usually, if someone removes content, you will see "not available anymore" or something like that.

I think it showed me that not believe really anything unless it can be confirmed. I am sure that isn't the only channel utilizing AI.
The one i watched yesterday was just audio with no video.
I am not saying that there was something sinister or bad happened but i feel i should have read the content before even starting listening. So if you don't do that you can get fooled thinking it is the person you believe will be taking. Even if its not the person but an AI instead.

Scannerman · Dec 15, 2025

coffeeaddicted said:
Well, i found another channel and its not him.
I think that the channel i watched was not him speaking but an AI from a transcript perhaps.
He doesn't have a YouTube Channel.

Usually, if someone removes content, you will see "not available anymore" or something like that.

I think it showed me that not believe really anything unless it can be confirmed. I am sure that isn't the only channel utilizing AI.
The one i watched yesterday was just audio with no video.
I am not saying that there was something sinister or bad happened but i feel i should have read the content before even starting listening. So if you don't do that you can get fooled thinking it is the person you believe will be taking. Even if its not the person but an AI instead.

View attachment 156997
View attachment 156998

Social media has a bizillion of these AI generated likenesses of celebrities from cartoon characters to presidents (including Trump and Biden). What techies should get on top of is ANTI-AI websites and start promoting those. Methinks there may be a demand for that sort of thing soon enough if there isn't already. Indubitably, AI will likely steal from Anti-AI web sites regardless, just as it robs artists, authors, and musicians.

gunrunnerjohn · Dec 16, 2025

Hazel123 said:
Maybe he just deleted it. That's happened before - or it didn't get saved in your history. I don't think AI creates false podcasts of real people

Not true, deepfakes are popping up more regularly on YouTube and social media. If you haven't seen at least one of these, you're not looking!

AK6DN · Dec 16, 2025

coffeeaddicted said:
Something happened. Not sure if that happened to someone else.
I am rather confused.
I was watching yesterday a podcast ( i thought it is one) that was from Michio Kaku about why we never can go to mars.
Now, today i wanted to watch another episode but it's gone. My history has no knowledge that i ever watched it.
I noticed a lot of channels that have Mitchio Kaku but seem rather strange. Perhaps AI generated. I have the feeling that what i was listening wasn't Michio Kaku but rather an AI that sounded like him.
This is so bad. Most channels i seen are most new. 1 day, 2 days. maybe a month. Not older.
I feel i am loosing trust in YouTube.
Anyway, it actually scared me.

So looks to me like a new thing is basically scammers taking the text of actual podcasts, using AI to generate a video of that person reading the text, and posting it on YouTube, etc to get clicks and advertising $$$ and thus make money from posting someone else's content. Expect to see lots more of this as AI gets faster and better at generating realistic videos. For any videos online these days (especially Youtube) take with a large grain of salt as real vs AI. JMHO.

Chat GPT - the good and not so good

Member

My Computer

Well-known member

My Computers

Well-known member

My Computer

Well-known member

My Computers

Member

My Computer

Well-known member

My Computers

Stock Market Wizard

My Computers

Member

My Computer

Well-known member

My Computer

Well-known member

My Computers

Well-known member

My Computers

Similar threads