LogFAQs > #974905859

ParanoidObsessive posted...

If you coded your hypothetical AI with that objective, it would certainly strive to achieve that as an objective - but the same could be said of human CEOs.

Well, yes, but you can't really code a human CEO with a new objective if you don't like their current one. Not without some very ethically questionable methods, at least.

ParanoidObsessive posted...

You're essentially adding extra stipulations to the role, then holding it against human CEOs for failing to live up to a standard that doesn't actually exist.

It's a standard they'd meet if they had no other choice but to meet it. As it stands, they choose not to, both because they have to compete against others who are making the same choice and because making that choice yields greater benefits for them, but by taking away not only that choice but also their ability to lobby against laws that would take away that choice (which is precisely why they still have the choice), things stand to look a lot better.

ParanoidObsessive posted...

A lot of the problems people see in modern corporate culture go back to the fact that stock investment and shareholding mandates a degree of growth beyond mere profitability. It's no longer enough for a business to be successful (ie, what most people think the purpose of a business is), corporations must strive to have positive growth at all times forever. Which lends itself to both anti-employee and anti-consumer policies, because literally nothing matters other than the bottom line.

The thing is, that's actually fine, provided there are sufficiently robust safeguards in place to ensure employee and consumer welfare aren't sacrificed in that pursuit. Because there aren't, those are corners that get cut in the race to grow the fastest, simply because that's what has to happen to stay competitive.

What's needed to stop modern corporate culture from being so harmful is simply a level playing field with inalienable standards that prevent those harms. AI can absolutely provide that, and spends a lot less time whining and suing people over it.

shadowsword87 posted...

I disagree with CEOs being optimal replacement, there isn't enough text to properly build up. Plus with enough random curveballs being thrown their way it can be weird.

It'd be a little awkward to train because so much of the history isn't simply text, but for the most part I would expect that an AI with records of CEO's decisions and contexts over the last couple decades would be enough to replace them pretty completely. The main problem there would be that there'd be nobody higher up to check on it and correct its course if something went wrong, and it would definitely fall flat in trying to deal with brand new paradigm shifts (though truly brand new paradigm shifts are exceedingly rare, such that a sufficiently trained AI would likely be able to relate any new one to enough old ones to manage).

shadowsword87 posted...

Middle management however, or a supervisor, or hell HR? Oh baby easy replacement.

Very much so. Front-line supervisors are still good to have for day-to-day directions and advice, but pretty much anything above that only exists to roll up performance metrics and compare them against some threshold.