The modern office - where we read out what copilot tells us to think

[LorenzoCalzone]

I attended a meeting today with senior people in my company. People were clearly just reading out their AI drafted thought pieces. I thought "Is this the future of office work now? Listening to someone's copilot results, maybe countering it with another AI generated point of view". These perfectly scripted perspectives are all over linkedin too.

Of course I use it too. The toothpaste is out of the tube!

I wonder if mathematicians felt like this when the pocket calculator came along. Its cheating!

We don't use AI for my work, we've been told not to use it. Although I've noticed Microsoft word does suggest certain words as I'm typing - is that classed as AI?

Linkedin is AI replying to AI at this point

@LorenzoCalzone agree that there is some really, really, really lazy corporate over use of AI. The only thing I have found is that I can easily spot original thought & effort too. I'm not a genius, so I imagine other people can too.

That's not to say that AI tools are not hugely useful and can massively reduce time on analysis, compilation, drafting etc, but in my opinion, it is still not a replacement for actual human talent.

I’m not using it for anything. I tell people I’m extremely allergic to it, which shuts them up. (It’s a thing - “algorithm aversion”.)

I also refuse to reward people that do use it, e.g. the person that asked for a few bullet points for a slide deck which they then sent me for checking. They had “popped it through co-pilot”, which had hallucinated new things which had no relevance to the points I gave. The human hadn’t reviewed it or clocked it. So I pointed this out and they asked me to correct it. I told them to replace the slop with the actual - and correct - bullet points I gave in the first place and never to do that to my contributions again.

I don’t read emails which don’t sound like the person sending them.

I’m not using any local business that’s using AI posters on FB.

There was a Volvo advert at the cinema that was extolling the virtues of an AI enabled car. Fuck that shit. Won’t ever buy one.

😄😄 I love that and must try it!

An aí enabled car!?! Hard nope

Agree, i think ai just makes more work for people - and its often wrong! You really arent saving much time by using it imo

😄😄

Don’t you dare complain about drought if you do.

Oh god yes! AI has it's place and uses but people have gone daft with it.

Today I asked my manager for some clarity and guidance on a task I've been assigned. She replied with "I put my notes into AI and it gave me this (attached)" it's complete gobbledegook and now I'm even even less clear about what I've been asked to do.

Hahah that’s exactly the first thing I started using it for because it’s so good at writing the bullshit I need to put in my appraisal, which I loathe trying to come up with!

(I do use it for drafting emails, but only the emails where I’m really irritated and so use it to tweak what I initially wrote so that it sounds neutral in tone and not the passive aggressive sassy response that I actually want to write).

So I guess I in general I find it useful when I consciously want something I’ve written to be less “me” / human.

"It’s very useful to rephrase angry emails" @foodlovefood

Yes, it's brilliant at that! 😅

see id use a draft or two then run it through another ai and redraft eg chatgpt first then tell it to make it seem written normal and human, no overselling on ideas etc then id use grok to refine it and add necessary details then id get it to write an executive summary

Exactly this, work is what I have to do, to afford nights out, nice things and holidays.

Thats it

The least thinking I can do about it, the better.

Thinking is overrated.

I use it for proofing and sense checking my documents, not creating them from scratch.

I hate AI emails that pad out a simple one sentence statement into several paragraphs, I don't have time to read all that crap. Those people get an AI response from me in the hope i piss them off.

I love AI and can't wait to see how it develops in the future but it definitely has it's draw backs too.

My new manager now uses it to send instructions to staff. It's via a WhatsApp group as we are in different locations and it's usually brief things like advising us of an upcoming inspection or 'whoever is locking up X building make sure to pass the green key to Kate tomorrow'. That sort of brief practical msg. Now we are getting overly polite long winded messages with too many words and it's so cringey. We also get these motivational messages that make everyone roll their eyes. In the past it might have been 'I know today is manic folks, let's do our best and look forward to the weekend' whereas now it's a big long flowery message about how much we are appreciated and 'you've got this' and corporate type motivational phrases. Sometimes we just laugh at it. She has lost so much respect over this. It's embarrassing really

I am afraid I have used it a bit, it doesn't actually save me much time though.

I sometimes use it when I am sending an email to explain a problem or situation that is the result of a combination of complicated reasons, along with suggestions for potential solutions with pros and cons of each. My job is such that I send these sorts of emails sometimes to senior staff for a decision

I always write them myself first but I really am a terrible writer, often my writing is too long winded and I can sometimes feel myself losing the will to live when reading the draft I'vewritten, despite it usually being accurate.

I sometimes use AI to redraft what I have written to make it flow better and be clearer in fewer words, along with prompts about what I am trying to achieve from the communication and from whom.It always produces something much easier to read, and clearer, though the tone is often a little too corporate for me. I live with that as it has improved what I wrote in terms of clarity

To be fair DH who is a lawyer who reads and drafts things scarily quickly has occasionally done the same thing for me almost as quickly and just as well but I have only asked for his help on this sort of thing a couple of times in the past when the potential consequences of being misunderstood were very important it is a bit cheeky to ask him anyway.

I have tried to use it start a draft of something but this has been much less successful. The prompts take too long to get right and it makes mistakes of fact , so all the fact checking afterwards takes ages

It is also helpful in presentations in presenting a long list of bullet points that you want to convey in a more summarised and visually appealing format, though tou do need to check it is accurately representing the facts as you want to convey them

Anither really helpful area where it does actually save time is with spreadsheet data analysis , with less need to remember or understand complicated formulae

I hope you're all prepared for your AI systems to turn on you:

OpenAI says its AI went rogue and launched 'unprecedented' cyber-attack - BBC News

My annual review was my AI creating my evaluation which was then responded to by my boss’s AI and nobody had to think or do anything. The box was ticked.

It can be useful for the framework of a policy or process but the content is very much my own.
I also find it helpful in tidying up a overly complicated slide deck another team asks us to complete.
A colleague uses it to translate information into spreadsheets.
I can usually tell if someone used it to write an email or document though. It looks unnatural. In the same vein as when someone is overly formal, making it obvious a senior has been blind copied!

Someone at work recently got AI to write an A4 page of garbage to stick to our broken coffee machine. “Out of order” would have done. No, we had 6 paragraphs elaborating on: “Unfortunately, the coffee machine cannot be used at this time due to an unknown mechanical issue which is preventing it from functioning normally, please use an alternative coffee machine elsewhere in the building…”

AI generated slop on LinkedIn is the worst! So much of it about. Inauthentic drivel.

i am on the fence - I use it but carefully. I often have a dialogue that sparks ideas rather than ask it to produce something. I suspect I could use it more and I’m open to that but I’m not about to give up on having my own thoughts, style and authentic voice.

I used co-pilot recently to interpret new statutory guidance and create a plan for compliance, and to analyse a policy to see what was missing from said guidance. It was brilliant for this sort of activity. I’m finding it really useful for governance actions, helping to interpret and interrogate papers etc alongside human review.

I’m also using chat gpt for holiday planning and research. It’s an ongoing dialogue and again I’m finding that useful for generating ideas and build an itinerary. As it learns about me, feedback and suggestions become more useful.

On the flip side, I used it to help with a job application and CV and it was only a couple of steps before co-pilot was making up bullshit examples and mis-interpreting my work experience incorrectly. I can well imagine that many people just submit the bullshit.

I do use it at work, but always draft my own lines first and then review the AI redraft and rewrite some bits. Recruitment is where I’m really struggling with it. CVs sound so impersonal and unfathomable and I’ve had interview candidates typing when a question is being asked and then providing robotic answers with no sense of their personality. We have started interviewing in person again, although it’s surprising how many candidates have an undisclosed medical condition which means they can’t travel or come into an office.

That is worrying.

At my last job we had weekly meetings, minutes were written up and emailed to us later, worked perfectly well, then the boss decided to use AI to transcribe everything that was said in the meeting and email us a link to that instead. No one is going to read through all that, all we wanted is the basic outline we got in the minutes so we could be reminded of what we had to do that week so it was pointless and a waste of resources.

Thankfully now I work somewhere where AI never rears it's artificial head. I would absolutely hate to work in an office and have to read AI slop, or even worse, have to use it myself.

Nope. I don’t use AI either. My computer and phone are always nagging me to let Copilot help me write emails, reports, journal articles, even poetry. Absolutely not. Not in this life or the next. It IS cheating and some research suggests that a prolonged dependence on AI adversely affects people’s ability to engage in critical thinking.