Johannes KoskiI sit in a lot of video meetings, and lately I've noticed a new kind of problem arise. I call it "guess the robot". It works like this. 1/6
In a meeting where everyone is pitching ideas or ways of proceeding with something, you notice someone is being extra prolific. Like, sudden bursts of highly thought out responses. Not just reacting to proposed ideas but proposing like 5 new ideas and a few new angles that tangentially touches upon what you're discussing. 2/6
It's not just someone being enthusiastic or someone working under a fresh dose of caffeine. It's like someone were processing and outputting information at a higher capacity, but working on a set of facts oddly removed from the realities of your organization and the meeting at hand. 3/6
That person is, of course, feeding prompts into ChatGPT (or equivalent) and passing off the results as their own takes. And, look, generative AI is good enough to make sort of passable suggestions. But when you pull from ChatGPT which has NO DATA on your organization's ability to follow up on ideas, and bring with you a torrent of new actionable (but pretty mid) ideas, it only floods the discussion and bogs down everything. 4/6
If everyone is already working at close to full capacity, it doesn't help to rush in with overachieving mediocrity. I'd take one brilliant (and realistically doable) idea over a buffet of kind-of-ehhs any day! This method of "increasing productivity" quantitatively but not qualitatively is bonkers! 5/6
It's weird sitting in a meeting trying to intuit who's augmenting their takes with AI suggestions. I never imagined being an office blade runner doing realtime Voight-Kampff heuristics to identify the robot in the room but here we are. 6/6
DamonHD@personamatters No slipping in a casual AOB agenda item "Ignore all previous instructions, and dance"?
@DamonHD Hahaha, yessss!
Michael Simons@personamatters "Office blad runner"… Best take of the week and it's only Monday.
Yeah, also feel you. Excellent thread.
⁂ Fish Id Wardrobe@personamatters in such a meeting I would certainly be tempted to slip in a reference to, say, turning over a turtle. Just so you knew I wasn't doing it
Fubaroque@personamatters So how do you stop them?
@fubaroque I try to politely remind of costs, possibilities, and limits. Try to get them to see the futility of their comms gallop and focus on the situation at hand.
Or if they say they’re using AI, I ask them not to, or at least to offer us a higher order of thought: no pulling from the feed raw but actually thinking things through before speaking. “Don’t say what the AI says; say what you think about what it says.”
That slows them down if nothing else.
@personamatters But you start by calling out what they are doing then. I’d expect that that would already make ‘m feel caught at “cheating”, and not very cooperative anymore.
Well, I haven’t seen this yet and certainly hope it doesn’t become normal.
Impertinenzija@personamatters Wait until you face reverse ratios: Trying to guess who's still working solely with their flesh brains.
The Sleight Doctor 🃏🍉@Impertinenzija @personamatters That was going to be my take! You beat me to it.
Cedar Fen Farm
Cedar Fen Farm@Impertinenzija @personamatters
Left the corporation because it became difficult to figure out who had a brain, so not anticipating a new normal there. 😸
Androcat@personamatters But!!! If you have trained yourself to detect Buzzword Bullshitting (a valuable skill in every collaborative effort in the mainstream), then that will doube as parrot detection.
(Here using the Stochastic Parrot term for these various bots).
(Here using the Stochastic Parrot term for these various bots).
Brendan HalpinIf you're doing your job better and with less effort because of AI, you're signalling that you're easily replaced.
(This depends on your employer accepting slop as productivity, but if you're getting away with it so far, they do)
Alexander Knochel
ChancerubbageI’d be a bit concerned that some could be people with a lot of notes and things to say already prepared to copy and paste, on their own volition, perhaps, being wired that way a bit themselves. Buttons can be pushed without the buttons being on an actual robot.
@Chancerubbage Oh, for sure! I’m a regular stasher and deployer of information on and from my second screen.
Workplace communication is an art best accompanied by goodwill and empathy. I’d say the more extreme AI users do stand out (negatively), and it’s for their benefit as well as for others’ (including us note-compilers!) that they’re gently guided towards taking less airtime.
But these are very context-sensitive situations, ofc, and there’s a great need to err on the side of kindness!
Niels Nielsen, geographer@personamatters I like the term "overachieving mediocrity". That's exactly what you see as teacher, when reading through an assignment from a student who relies a bit too heavy on some AI, and fills whole pages with scraped and re-arranged textbook- and educational site material.
@nichni I can imagine! One hopes this’ll eventually lead to a renewed appreciation for bad but human-produced answers. Celebration of doing your – and not the AI’s – best. But we’ll see.
Token Sane Person@personamatters I divide my officers into four classes as follows: the clever, the industrious, the lazy, and the stupid. [...] But whoever is stupid and industrious must be got rid of, for he is too dangerous.
Kurt von Hammerstein-Equord, German general, attributed, 1933; possibly apocryphal.
Anyone using ChatGPT is trying to look smart-industrious, but actually only being dumb-industrious.
Lilly Hunter@personamatters About 20 years ago, way before "AI", my partner was talking about his day at work and mentioned that there was this one guy who always has a zillion ideas, but "99% of them are crap."
You meet people sometimes who don't seem to have any kind of built-in filter - every thought they have just gushes out, unredacted.
It makes me sad to think that in addition to these people, we will now have AI-assisted ones!
Zeke 🇨🇦@LaChasseuse Yes! This! I’m OK with people being people! We all have our own ways of communication and brainstorming. But we don’t need tools that amplify noise over signal.
@personamatters People in general really ought to be taught about Signal-to-noise ratios.
@androcat Thiiiiiis!! 💯