Re: Mind's Eye Moral Enhancement

I'm not sure it has to do anything much to us Allan - though potentially it changes everything.  The machines could soon be biological - they can already record information as DNA.  Corrupting programs might be stopped by surveillance routines.  We could look at this as human, even soul enhancement and as educational.

On Monday, 2 March 2015 07:35:34 UTC, Allan Heretic wrote:

AI sounds cool.. several problems though it would be easy to program violence in, the manipulation show with out a chip is going to suddenly change with a chip.  RIGHT!

The other problem is the soul..  and the mix or no soul  pure AI will it contain a soul?

تجنب. القتل والاغتصاب واستعباد الآخرين
Avoid; murder, rape and enslavement of others

-----Original Message-----
From: archytas <nwterry@gmail.com>
To: minds-eye@googlegroups.com
Sent: Mon, 02 Mar 2015 8:09 AM
Subject: Mind's Eye Moral Enhancement

Humans developed to live in small communities - we were pretty murderous in them and you now are exposed to only a tenth of the chance of dying a violent death.  We are not well-equipped for today's global circumstances.  He are not much good at large scale collective moral problems.  Moral enhancement in traditional form has been about education, religion or short term drugs and lobotomy-type intervention.  Artificial intelligence is another possibility.

Far from proceeding in the rational way set as an ideal, most of our moral views and decisions are made on immediate intuition, emotional response and gut reactions. Reasoning, if we do it at all, is often just rationalisation of what we intuitively thought anyway. To overcome our biological and psychological limitations, we could develop moral artificial intelligence.

Many are very scared of this, perhaps because they know they are not strong moral agents.  Some think such machines would recognise us for what we are (a danger to the planet) and kill us off.  Given our potential to do this to each other, I'm dismissive of the machine problem.  MIA could monitor a lot more than we manage as humans and point out personal bias and advise on the right course of action according to human moral values.  Agent-tailored MIA would preserve moral pluralism and help the individual's autonomy by removing the restriction of her psychology.

I have volunteered Gabby for the first MIA chip (no wait, that was Cartman with the V chip in South Park).  In fact, AI is a;ready helping with a lot of learning.  We are introducing AI into fraud management systems with patents being filed - http://www.freepatentsonline.com/20150032589.pdf - car driving, medical and dental analysis, narrative generation in entertainment - http://eprints.hud.ac.uk/23153/1/118.pdf - Big Data will drive Big HPC and Complex Analytics. Supercomputers of the future will need to: (1) Quantify the uncertainty associated with the behaviour of complex systems-of-systems (e.g. hurricanes, nuclear disaster, seismic exploration, engineering design) and thereby predict outcomes (e.g. impact of intervention actions, business implications of design choices); (2) Learn and refine underlying models based on constant monitoring and past outcomes; and (3) Provide real-time interactive visualization and accommodate "what if" questions in real-time. This will require an evolution in algorithm and system design, as well as even chip architectures to manage the power-performance trade-offs needed to attain a new era of Cognitive Supercomputing.

Heads in the sand on this folks?  Or would you have the "implant" like me if one was available?

--

---
You received this message because you are subscribed to the Google Groups ""Minds Eye"" group.
To unsubscribe from this group and stop receiving emails from it, send an email to minds-eye+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

---
You received this message because you are subscribed to the Google Groups ""Minds Eye"" group.
To unsubscribe from this group and stop receiving emails from it, send an email to minds-eye+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.