On Using AI at Work....

in #ai15 days ago

I regularly use ChatGPT at work, not to write content, but to 'improve the grammar of the text below'.

I literally just type in that phrase and upload the text I've just written.

And it doesn't matter what correspondence I use this for I always get an improved version with a more professional feel and better flow compared to what I'd written myself.

I use chat GPT with a subscription, it's pretty clever. Today for example I wanted to say something along the lines of 'your input helps us ensure legislation is sped up', which is clearly nonsense, but it managed to rewrite a garbled long sentence into two for me, and it sounded like what was in my head but i couldn't quite get out at the end of a working day.

I think it could end up making me lazier, a bit like a satnav!

NB using AI has been recommended to us - and I know another caseworker uses it because early on in this role I asked the AI to improve the grammar of some text, part of which had been written by a colleague, and that bit came back largely unchanged, as she'd already put it through the AI!

And we're not the only ones using it!

Screenshot 2025-02-04 at 20.20.01.png

According to this article half of all knowledge workers use personal AI tools at work.

Many of them do it on the down low as getting formal agreement to use non-approved work AI tools can be time consuming, and I get the feeling managemenet are happy to turn a blind eye because they are required to be guarded about the use of AI - to use anything outside of the organisation risks your data being used to train that AI for free, which could put your firm at a disadvantage.

Just remember GDPR etc...

Personally in my role that doesn't bother me, all I have to do is be careful to redact any identifiers, name and address, basically, when I'm using it, or if it's a really unusual and highly personal case I wouldn't enter that either.

And I think the same goes for many people working in the knowledge sector, AI can be a real boon, and you're probably at a disadvantage if yer not using it.

We just need to be careful that we don't get lazy, that is a risk!

Posted Using INLEO

Sort:  

Seems AI will be more skilled than us in the future.

Actually, using AI as okay as long as we use it for good thing and not intens to harm others. AI is develop to help use to do something.

I don't use that stuff much, but then I don't write much text. At work we are not supposed to use the external 'AI' systems to avoid leaking company data, so they have an internal one that can be used for tech queries.

I think we need to keep real human interaction or you will get cases like a cartoon I saw where someone generates an email from a bullet point and the recipient does the reverse.

I think that later case probably already happens, I mean I'm a bot, after all!

We may already have bots conversing on Hive. Sometimes it's hard to tell if an account is human.

Laziness is a slippery slope. I'm just thinking of the automatic spell checkers that I've been using since college. Far from improving my spelling, I often feel dependent on them. Nowadays I keep using translation tools to go back and forth between the languages I actually speak. While the final result is always a word or expression that I already know, before typing it in I could not come up with it myself. Now with Chat GPT (and lately DeepSeek), when I tell them to paraphrase something I wrote, it always ends up sounding much better. In fact, let's see how it paraphrases what I just wrote:

Laziness can lead to a gradual decline in skills. For instance, I’ve relied on automatic spell checkers since college, but instead of enhancing my spelling, I’ve become dependent on them. These days, I frequently use translation tools to switch between languages I already know. Although the output is always a word or phrase I’m familiar with, I often struggle to recall it on my own before typing it in. Similarly, with tools like ChatGPT (and more recently, DeepSeek), whenever I ask them to rephrase something I’ve written, the result always sounds significantly better. In fact, let’s see how it paraphrases this very text.

I wonder what would happen if you just kept on inputting! That final one did sound very AI!

If you made it paraphrase the paraphrase, and just keep doing it over and over again... I suppose after a couple of version it would reach the limit of language and use the same expressions again.

Do you think the AI version I posted sounds artificial? Honestly, I could´t tell.

Maybe mix it up next time, see if I can tell!

As for grammar, It is good at this as well using Chatgpt in codes is very fun. I wish I was a programmer. It would have been great. Just mention what you want and it will be converted into any code you want.

I think that grammar checking is a perfectly valid way to use AI. Though I haven't tried it myself, yet. But I do know that using earlier grammar checking tools helped me to learn the rules easily. So I never felt that made me lazy.

I'm not sure why, but reading your post made me wonder if politicians have an Artificial Stupidity tool for their speech writing?

I'm pretty sure it's standard practice to dumb down speeches for the masses. Before AI this was probably done by professional propaganda wordsmiths. I once read a speech by 19th century frontiersman turned politician Davy Crockett. It was so complex, so elaborately convoluted, composed of run-on sentences addressing various ideas at the same time, that it was hard for me to imagine how he could even deliver it, let alone what impact it would have. Supposedly it was well received, but I'm certain it went over the heads of most potential voters.

I can imagine they do.I don't think it's a great time for writers of any kind!

I am using co-pilot at work. Which is the Microsoft version

Microsoft keeps wanting me to switch it on, I resist!