[ExI] The Real Threat of Artificial Intelligence

[BillK]

On 26 June 2017 at 17:17, Stuart LaForge  wrote:
> There are several considerations about this situation that give me some
> cause for optimism:
>
> 1. AI are software-based therefore it will likely not be possible for any
> given company to monopolize the use of the AI for its own benefit. I mean
> if the NSA can't keep Stuxnet secure then how would Company X keep its AI
> out of the hands of competitors and ultimately script kiddies? Therefore
> the gradual dissemination of the AI would help ameliorate the inequality.
>

The AI that the author is worried about is the initial, simpler AI
that just automates production in factories and offices. For example,
software like Amazon is implementing to automate running their
distribution centres without human intervention, ordering by computer
with collection and delivery by robots.

This automation will spread to all factories and offices, with
lower-grade employment being the first group affected. So corporations
will become richer, with few employees. This software will be of
little use to the large number of unemployed or script kiddies.


>
> 2. It might be hard, without some heavy-handed top down management by
> government, to protect even the highest paying jobs such as CEOs,
> banksters, hedge fund managers, insurance underwriters etc. from
> automation. I mean if AI can make more efficient resource allocations than
> e.g. Jaime Diamond, what share-holder in their right mind would pay his
> outrageous salary and bonuses?
>

Top-level jobs are much fewer in number and will require much more
advanced AI to replace. This will come in later years, but mass
unemployment will arrive first.


>
> 3. The public sector might benefit from AI as much as the private sector.
> An AI might be better suited to determine optimal tax structures and set
> interest rates for example than congress or the Fed. Hell an AI might be
> better suited to congress itself since it could parse a thousand page bill
> in mere nano seconds in leiu of a congressman who typically can't be
> bothered to actually read it before he votes on it.
>

The public sector will be the cause of great unemployment as the
mostly paper-shuffling jobs will be automated throughout national and
local government. Again, the few top-level political jobs will be the
last to disappear.


>
> 4. As AI gets more generally smart, it would start to recognize that
> profits will be maximized for Company X if consumers can actually afford
> the goods and services provided by Company X. Therefore it might hire
> humans to do busy work for Company X just to be able to have a customer
> base for Company X.
>

Hiring humans for 'busy work' is pointless. Production will be much
cheaper, so prices will reduce drastically and become more affordable.
If the government gives everyone a basic allowance (funded by company
taxation and taxing the few remaining high-paid human workers), then
the population should still be able to buy the much cheaper goods.
This enables human survival, but doesn't provide 'meaning' to human
life.


>
> That's about all I can think of for now, but I think you get my general
> drift. I think AI will be just as disruptive for 0.1% as for the masses
> and those that benefit the most will be the most flexible and adaptable
> and not necessarily the wealthiest. Just my two cents.
>

Being flexible and adaptable is useful at present, but in a future
where automation runs everything? Giving all the unemployed humanity a
reason to live will be required when they can survive without effort.
Society will have to be restructured. Entertainment, caring for each
other, games, sports, and yes, many will become couch potatoes or VR
addicts, drug addicts, etc.
Oh, and a large police force will probably be required (with robot
assistance) to keep the unemployed from creating too much havoc.

It will indeed be a Brave New World.

BillK