We now have our own bespoke company artificial intelligence interface called Alan or Al for short. The folks in IT keep sending emails telling us to ask Al about anything. As always around here, statistics are the root of all evil. Someone has spent a lot of money on Al, and we need to justify that cost by showing how productive it is making us. I know this because I discovered that any questions asked cannot be deleted because they are being used for tracking purposes. Note to self: do not use the work AI for any personal questions.
I have done the training on creating AI prompts, but unfortunately it was aimed at client-facing teams so very little was relevant to my job. Still, not to be outdone by new technology, I will embrace Al whether he likes it or not.
Effective AI?
Big Bad Boss has asked me for a summary of our share plans and their tax treatment, so I ask Al. As suggested in the training, I am very specific with my request: ’please give me a table of equity plans including stock options, restricted shares and restricted share units and their tax treatment in the UK’. You do not have to say please but it feels better to me to be polite. Al takes only a second to give me a list just as I asked laid out in a beautifully formatted table. Gulp. Al has done this rather better than I would myself in a percentage of the time. A sense of foreboding and panic washes over me. Big Bad Boss could start asking Al directly and save on my salary altogether. This is scary, and I am not ready.
A closer look at the table, however, reveals an important fact: the table is wrong. Wholly and specifically wrong. I know there is a difference in tax treatment for restricted shares and restricted share units, but the Al’s table has everything listed as exactly the same. I wonder if my knowledge could be out of date, so I contact my friend in the tax department. He confirms my suspicion; the data in Al’s table is incorrect. Gulp. He gives me a link to a website which lists all the tax treatments internationally, so I set about amending the table with correct data.
Al has helped me to present the data in a useful table and that has saved me quite a bit of time messing about with formatting, but I am still worried about the inaccuracies. How many of our staff would not think to check the details. Take my colleague Lazy Susan: I mean she does not think much at all really.
Artificial communications
The other place I am seeing artificial intelligence creeping in, is in our email system. Where it used to underline the odd typo or duplicate word, now it is constantly underlining perfectly valid words I have put in for emphasis. When I click to find out why things have been marked in read, it demands ‘conciseness’, suggesting I reduce everything to words of one syllable to make things clearer for my reader. Well, I think, if my readers cannot ‘actually’ manage the odd word for emphasis, no amount of clarity will help them. Sometimes, I select the alternatives suggested, but the result is boring, flat, and monotone. Sadly, I have noticed my colleagues seem to be more easily directed and many of the emails I receive are dumbed down to the level of a toddler book.
Not only that, but ordinary salutations in emails are changing again. I got used to everyone using the American ‘hi’ rather than the long held British ‘hello’, and I now barely pause when someone uses ‘hey’ even though it still feels a bit overfamiliar at work. Now, some emails are coming in without any salutation at all, and certainly no further nicety of ‘did you have a nice weekend’ or the like. No, they go straight to the point with a demand for data or whatever it is. Signatures are even worse. No one signs off with ‘kind regards’ anymore and ‘thanks’ is only used in a subtly hostile way, as in I-expect-you-to-do-this-by-end-of-day-today-thanks. Everyone is so concise that they simply default to their email signature with full name and job title, along with our corporate logo, not even signing off with their first name first. I will not join in, but I am nervous of using ‘kind regards’ now it has become unfashionably lengthy. I stick to a bland ‘thanks’ and hope I do not sound too passive aggressive.
Big Bad Boss shares a note from the chief people officer about changes to the hybrid-working policy and how these will make us all more creative and productive, referencing links to studies on operational effectiveness and employee engagement. There is something about the formatting I recognise. Sure enough, I can replicate the note almost exactly using AI and, because I am looking for one, I find a mistake in the references. It is bad enough knowing how much corporate propaganda the Higher Beings, our executive team, put about, without them employing an AI tool to add to the chaos.
I point out the error to Big Bad Boss and tell him about the factual errors on share schemes data from Al. I know he will not do anything about it; it is career suicide to question the use of technology when you work for a tech company. No, we agree to continue to use Al to draft communications and set out data, but to religiously check any factual data against other sources. Hopefully, Big Bad Boss is savvy enough to realise my experience still counts for something, and that means that Al is not coming for my job anytime soon. But then, why would Al even want a job like mine?
Next time…Candid looks at preventative healthcare