[ExI] AI and jobs
Adrian Tymes
atymes at gmail.com
Thu Jan 8 12:17:10 UTC 2026
Problem is, many of these jobs are seen as so critical that for AI to do a
demonstrably enough better job that most people would accept that AI can do
the job better than humans may be an impossibly high threshold.
The problem is not getting them to do better in simulations, or what you
would consider objective tests. The problem is getting them to the level
of proof that most people will accept, without having them actually do the
job before most people would allow them to do the job.
The likeliest solution is a very slow - by your standards - ratcheting up
of responsibility. Robotaxis sharing the road with human drivers for a
long time before humans no longer drive. Robo-mayors for a few small towns
well before AI heads of state. Et cetera.
On Thu, Jan 8, 2026, 3:34 AM Ben Zaiboc via extropy-chat <
extropy-chat at lists.extropy.org> wrote:
> It occurs to me that we are obsessed with the jobs that humans currently
> do being replaced by AI, to the detriment of humans, when it might be
> better to focus on the jobs that humans have proven to be bad at, where it
> would be much better if suitable AI systems could take over, to everyone's
> benefit.
>
> If we use the main principle of 'reduction of harm', mostly to the people
> affected by the job, not just to the occupant of the job (which should be
> secondary, when you think about it), I'm coming up with some interesting
> and probably controversial candidates for jobs that SHOULD be done by AI,
> simply because humans are so poor at them.
>
> Here's my list:
>
> Heads of State (in a general sense. This includes any individuals or
> groups of people who actually run a country)
> Military, all levels
> Heads of Police
> Judiciary
> Economic & Financial Policy
> Education Policy
> Communications, all levels
> Transport, all levels
> Healthcare, most roles
>
> With secondary principles of efficiency and fair universal access to the
> benefits that each role provides to the public, I think it just makes sense
> that AI, once it's capable (which is very probably not far off), should do
> these jobs, and in fact that humans should be barred from doing them, given
> the harm they can, and often do, cause. If there's a case, as many people
> think, for barring humans from driving vehicles on public roads once AI can
> safely and efficiently do the job, there's certainly an even better case
> for barring them from these much more dangerous jobs.
>
> I've probably missed some, please feel free to add to the list.
>
> ---
> Ben
>
> _______________________________________________
> extropy-chat mailing list
> extropy-chat at lists.extropy.org
> http://lists.extropy.org/mailman/listinfo.cgi/extropy-chat
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20260108/db8c4c86/attachment.htm>
More information about the extropy-chat
mailing list