Not just burnout, opportunism features with several users I’ve spoken with. The level of ignorance surrounding ChatGPT is staggering.
One egregious use I know of was a developer who used it to write software to analyse a government dataset despite their department having put in place specific and targeted restrictions specifically against any such activities.
Their workaround was to use their private email to exfiltrate data and subsequently introduce the code.
Their rationale was that it didn’t harm anyone and their ICT department would vet any code. They were not concerned about this private data showing up on the ChatGPT public log, nor were they concerned about the accuracy of their code.
I think that this is just the tip of the iceberg and I think it’s going to take a serious data breach of identifying information before people lose their jobs over this type of misuse.
People have already lost jobs for this very behavior back in 2022. I remember reading news about Samsung managers even being sued by their employer because they fed secrets into ChatGPT. And if I remember correctly this serious breach was committed for the sole purpose of brushing up some emails.
What is the ChatGPT public log?
For a period the interactions you had with ChatGPT were public and a live stream was available.
At the time when I looked at it, there was an astonishing amount of non-english traffic, but that might have been due to the fact that my UTC+8 timezone in Perth is the same as mainland China.
I had a quick search just now to see if I could find a link, but all I can locate is posts about new privacy controls, so perhaps that "feature"went by the wayside at some point.
Could that have been a public live stream log of a different provider’s model?
I’d be astonished if that was OpenAI’s. Can’t find any articles, threads, or screenshots after extensive research (a minute of web searching). And didn’t hear anything at the time going back to December 2022.
I’m fairly certain that it was ChatGPT, but I’m going from memory. I have a hunch that I saw a Hacker News show and tell post.
Update: It was in my bookmarks.
https://www.kaggle.com/datasets/noahpersaud/89k-chatgpt-conversations
Oh thank you! Very interesting.
Hope users of chatlogs[.]net (now defunct, spam redirects) realized they were conversing publicly!
That’s… not true at all.
Someone made an app using the OpenAI api, offered it for “free”, and scooped a bunch of users conversations (sort of. 89k isn’t that large) before getting shut down. That’s not at all what you were alleging.
Don’t leave me hanging. Was the dev fired / sent to jail?
Yup. Throw in multiple levels of contractors in various counties and guaranteed AI is being used. (Think subcontractor’s employee using AI to keep up.)
AI is a great spying tool.
Of course burnout is a real thing and most of us are increasingly overburdened. Most work just seems to suck more and more and the lines between personal life and work life keep blurring. Nonetheless, this AI use would happen with or without burnout - it’s just opportunism. Also, Who the hell is supposed to do this training the article refers to? Does your office already have an AI tsar that could train people on best use? I haven’t seen much of that yet.
The headline sounds a bit like “Burnout is pushing construction workers to use nail guns.”
… according to a work trend index published Wednesday by Microsoft
Yeah, I’m going to go out on a limb here and call bullshit. No one is turning to AI to alleviate burnout. The only tasks these LLM tools can reliability accomplish aren’t worth using an LLM for.
Patently not true
Not a day goes by where I don’t use our company’s internal LLM instance to generate or debug some code. It isn’t due to burnout, it’s due to convenience.
