Sometimes I encounter posts here which are nothing more than copy-pastes of what AI had answered to a particular question (maybe wrapped between about 20 words about how it was asked of AI).
I think this is indicative of the person posting such AI drivel knows absolutely nothing about what they’re talking, while faking being useful.
Then there’s the copyright issues with this AI stuff. There are court battles happening everywhere about all kinds of AI related stuff. Itsfoss could run into legal issues if they are found to host AI generated content. Basically: it could be a legal minefield, or it could be a non-issue and anything inbetween.
I personally find people posting AI generated stuff intellectually lazy and in poor taste.
Am I the only one who takes issue with this?
Basically I take issue with people letting AI generate their post.
I agree on that point, think, lije most if I see a question on this site I check with google to look for answers, see if others are having similar issues, see if there is an answer … etc. If I find a site with images then I may link to it to explain how to arrive at a solution. But I worry about AI and the results offered, as I have said before glad I have stopped teaching or marking homework as I may not agree with the source or results.
I disagree. I think if it is being used to address a computer command issue, it is about the same as posting the contents of a man page or a help document.
After all AI is just a big encyclopedia and can be an efficient way of retreiving information.
I find the google AI answers helpful, and sometimes I post them as a quotation.
The way you phrase a google query seems to affect whether you get an AI summary or a list of links
I dont think there are any copyright issues… at least not for computer queries.
I think if one quotes an AI reply , one should make it clear what the source of the quoted material is.
I have steered clear of all things AI until this week. As I understand it so far, there are different AIs the way there are different search engines? I was on Bing and kept seeing the top area as the answers to my search questions, kind of like a summary of all it had scanned the web and found. Of course, each answer or part of the answer referred back to the source website. Okay. Fine.
But I decided to actually select Co-Pilot along the top edge where you either choose All, Images, etc. Then I began asking questions on my new computer build I was researching. This was cool, I thought. It’s like asking which motherboard will fit in my current case and get a few options to check out.
When I got down to a BIOS question, I asked about how to flash it. The resulting answer, in steps (supposedly in the correct order) had me download the zip file and extracting it to a USB drive. Then I was to format that USB drive to FAT32. Huh?
So I kind of reiterated the steps given, by saying, “So I extract it onto a USB drive and then format to FAT 32?”
The AI said, “Well, actually you should format the USB drive first or else you will lose your extracted files.”
Ya think? That made me doubt everything in this conversation. I am thinking it scans multiple articles but it is not quite capable of analyzing the data it finds to ensure the steps are put in the right order.
So I think one can probably use AI to narrow down search results. But I don’t think I would rely on the responses without first checking other sources.
I have a huge issue with AI doing “creative” stuff - i.e. art, music and impersonating voices… it’s not creating anything - it’s stealing / plagiarising… recent case in point, Scarlett Johannssen was reported as seeking legal action, after some company asked to use her voice in some marketing, and she turned them down - they went ahead and did it - using an AI simulation of her voice - that’s creepy and I think evil…
ALL computer generated voices or spoken words should be like Stephen Hawking’s computer voice - so we know…
I don’t so much have an issue with using it to resolve technical issues - but - I wouldn’t use it myself.
I think it depends.
It seems to be good at explaining computer jargon and how things work.
I dont know about precise instructions for doing something… it can get things muddled … like in @Sheila_Flanagan 's reply, so you cant follow AI instructions without thinking about the consequences.
Yeah, I use it like an extended set of man pages, and
I use it like a searchable Linux Handbook, and
I sometimes post its replies when I cant find a suitable link… I think links are generally more authoritative… but there isnt a link for everything… AI will attempt to synthesize an answer to an unusual question… it may make mistakes, but people make mistakes too. Sometimes it can get off topic and answer the wrong question.
I dont see anything wrong with posting an AI reply… as long as it is identified as such.
Oh dear, isn´t this putting it a bit harshly…
… or at least some kind of over-generalization?
I might be wrong, but at least for me it´s like this:
If I know the answer to a query/problem or at least I think I do then I go ahed and post it. Thus trying to provide some help.
If I don´t know the answer I´ll have a look at the man-pages or I´ll try to find out anything relevant on the net.
If I can´t find anything worth posting within a reasonable period of time I might consult ChatGPT or perplexity or whatever.
In this case I mark the answer as AI-generated. One should at least be honest about things like that, right?
In most cases I add a disclaimer like “Take it with a grain of salt” or something like that.
No need to mention that some personal additional investigation might be required.
I think adding some AI-related stuff from time to time shouldn´t be regarded as objectionable as long as it´s marked as such.
Plus: as @nevj already said:
So, wouldn´t we miss out on potentially good or interesting stuff if we excluded AI in our answers in general?
It´s just another way of researching.
I guess most of the reservations about using AI are derived from the fact than many people tend to use AI “the wrong way”:
They like to delegate their responsibility to AI, which means they don´t want to do any research themselves and as a result they are offended if part of the information generated by AI are incorrect.
Rather than going down this road one should see AI as some kind of stepping stone to his or her own research.
Some people are often happy to know where to start from instead of “looking at a blank sheet of paper” for a considerable time and slowly getting frustrated.
Just use AI responsibly and I think it has its justification here as well.
So what do you think of the Google AI summary that appears at the head of some searches?
I have quoted those… it saves giving a heap of links. Would you prefer the links?.. or maybe I should just give the google query phrase and tell people to google it themselves?
Some questions require a large amount of writing to answer and I would be reinventing stuff already known to write it myself.
I usually refrain from topics I do not know… unless I can flag what might be a helpful link. You will never see me in audio or video topics.
Some topics are more opinions than information. There is probably no role for an AI response there.
Does google always give the same answer to the same query, regardless of who submits it? AI such as perplexity definitely does not… it learns from each users history of queries… so I cant say to someone… feed this query to perplexity… I think I can with google… but correct me please?
My wife is currently using chatgpt … what a waste of time and effort.
Perhaps 30 years back she would have gone next door and asked the old guy who lived in the village about what happened in the village before the war, 20 years ago she would ask AOL or jeeves, 10 years ago google. But now she spent a morning on this new tool.
Will she be any better off for the research, I doubt it.
But it appears so easy to use and fihd things. Perhaps our site will stop existing in the long term and we will all be replaced by these tools.
I tried to get ferry timetables on it but in the end it could not answer me, I had to mail the agent. Its not that wonderful
It is all about having access to information.
Power has shifted towards individuals because they can now access information more easily and do not need to consult professionals for as many things.
But there are limits
many library books are not digitized… google tried and was prevented by copyright laws
individuals vary in their ability to understand and utilize information, partly because educational opportunities vary
I was recently forced / compelled to do some bullshit CBT about AI by my employer - needless to say - I was dragged kicking and screaming…
I really wanted to claim STRONG deeply held ethical reservations about AI…
And then they go and have late night sessions (late in my timezone) about how concerned they are of “neurodiversity” - I know I’m pretty much “neurodiverse” - my eldest daughter was diagnosed as autistic as an adult - and both she - and my mum - reckon I’m high on the spectrum… and my youngest daughter got an ADHD diagnosis as an adult.. Despite Autralia’s “stellar” (compared to the USA) health system - getting an autism spectrum diagnosis is still out of my budget (and dental is not covered here either - I’ve barely got any teeth left at 63)…
So - I abhor the whole IDEA of AI - and who thought it was necessary?
Solve proper science issues with it - DON’T REPLACE HUMANS! Take the drudgery out of research - but don’t take the fun part out of it…
And - same with crypto mining - some 3-4 years ago - some equated the amount of electricity used by crypto mining was equivalent to the power use by the whole country of Argentina! I’ve read that AI is WORSE than crypto mining at power consumption…
FFS! Musk’s Tesla are doing trials of self driving cars here in cities like Melbourne - FFS! Driving cars is fun sometimes too!
I agree , but remember all we see of AI when we do queries is the language model.
That may be the only useful aspect. A better human interface is something worthwhile.
How it looks up info is hidden , and varies between implementations. If it uses your query history, different users may get different answers to the same query. How can you use that for research… it is not repeatable if it tailors responses.
