Well said. But hopefully no one ever tries to help me with an AI dump. I really dislike this. 
As others stated, I’m here to communicate with humans, nothing else.
6 Likes
It solved itself. He has said goodbye.
There is still another offender… he/she has been warned.
AI generated writing styles are not acceptable either.
5 Likes
Can you share how to flag a post? Or maybe I just don’t have permission yet, being relatively new.
1 Like
3 Likes
+1 for a requirement. AI is not there to do all our work for is; rather, as a companion to questions. And I’ve already seen forums that are fully covered in AI slop and bots. (Of course, not our lovely Community forum!)
My opinion is that AI is not neede much here, except for providing answers. If everyone starts asking AI, what’s the point of that? And wothout knowing, a robot will be answering to you, while thinking that that user is a Linux pro.
I myself only use Ollama AI to explain some error messages while coding. I think AI can be helpful, also when posting to the forum with remark, but it also has many, many risks. If you see how much a specific OS (namely la fenêtre) is moving away from human focused to robot focused (Not for Daniel: Nothing against Windows itself, it’s a great OS, but full of AI slop.) If we stop working, what will happen? We were meant to work, to do. If we sit on a screen typing and reading answers, that doesn’t help to learn new skills.
I think this forum is one of the lovely places left that have little to no AI. And if people use AI, they should put a remark so people know.
4 Likes
Just found this in my inbox. It explains pretty good about the dangers that AI and their platforms bring.
3 Likes
Anything you submit to AI in a request is public domain.
So linking AI requests into a system in any automated manner can compromise parts of the system
That is what your example demonstrates.
3 Likes
It‘s sad that many people don‘t take data protection seriously. In today‘s world, pretty much amything is possible.
4 Likes
I am not defending OpenAI here but playing devil’s advocate…
There have been countless breaches similar to this for years. I’m not sure this is much different. It’s the first one I’ve seen that mentions any relation at all to AI.
5 Likes
I agree, to big of an issue is being made of AI!! For the record, I have asked copilot about several issues, about LFS, and the answers I have gotten are rather accurate, more so, than what I can get out of any human!!! So, for myself, copilot, is a helper tool, but like anything else, it can be abused!!!
4 Likes
Blaming AI may become the latest excuse for not doing your data security properly.
4 Likes
Yes, it an be accurate, but it can also misunderstand the request and give a nonsense answer. You have to be careful phrasing the question.
What I dislike about AI responses is that they are like a review article , but with all references to sources omitted. You cant tell where AI sourced its information. That is bad practice. No scientific journal would publish a review article without references.
In a lot of cases, your AI reply is simply a copy of wikipedia. That is plagiarism. At least Wikipedia gives references.
2 Likes
And I could care less about where, copilot, sources it’s data, as long as it is accurate, it comes with W11, and I use it and copilot, is a lot more accurate, than Google in AI mode!!!
Been cleaning an old Gateway PS/2 keyboard today and asked copilot for a teardown video, and it found a you tube video of what I needed!!! The old keyboard is now looking and working really well!!! I am now able to type well, the new USB keyboards are so thin, and I can hardly see the keys!!!
4 Likes
That is the point. In scientific literature, the way to check if something is accurate is to check the sources.
I think that applies to AI too. It is getting away with loose, unsupported statements.
It is more like a conversation than a proper report.
2 Likes
I am not doing a scientific experiment, I just need an accurate answer to an issue!!! Cannot see the issue with obtaining public knowledge!!! It is available, use it!!!
1 Like
The way to know you have an accurate answer is to be scientific… check everything.
I know you normally do that. … ie you dont assume that an AI reply is always
correct.
You just stated that copilot is more accurate than google. How did you check?
2 Likes
This is from the Australian Government
“In guidance to developers and content creators, the federal government has advised that AI content should be “clearly identifiable” by including labels that note content is AI-generated or embedding information to trace the origins of content — a process known as watermarking, which is more difficult to manipulate or remove than labels.”
They are mostly on about AI faking of pictures of people, but it applies to text replies too.
Apperently Google watermarks all its AI output.
4 Likes
Through trial and error!!!
1 Like
Yep. That is the same way scientists check… they call it experiment.
2 Likes
Really? Maybe I’m using it wrong… I ask it how to install flatpak on Fedora Server and he said I must go to Start and open the Microsoft Store?
2 Likes