|
AI researchers are warning developers to focus more on how and why a system produces certain results than the fact that the system can accurately and rapidly produce them. It's a system of tubes. Of switch statements.
|
|
|
|
|
And that's why in my first job AI was explicitly forbidden: people's lives were on the line and you had to be capable of explaining every single incident.
GCS/GE d--(d) s-/+ a C+++ U+++ P-- L+@ E-- W+++ N+ o+ K- w+++ O? M-- V? PS+ PE Y+ PGP t+ 5? X R+++ tv-- b+(+++) DI+++ D++ G e++ h--- r+++ y+++* Weapons extension: ma- k++ F+2 X
|
|
|
|
|
This is why we need to teach our AI systems human ethics and morals. Without this we'll end up with Terminator, Dune's Omnius, and Microsoft Tam (Tay?). Only with human ethics and morals embedded deeply in our AI offspring will we be able to coexist with them.
|
|
|
|
|
How would you teach ethics and morals to a large table of data-points that have no meaning? Really, AI as intended these days is a simple correlation between variables calculated automatically to fit the data given as input. There is no intelligence, it's truly a multidimensional best-fit algorithm.
GCS/GE d--(d) s-/+ a C+++ U+++ P-- L+@ E-- W+++ N+ o+ K- w+++ O? M-- V? PS+ PE Y+ PGP t+ 5? X R+++ tv-- b+(+++) DI+++ D++ G e++ h--- r+++ y+++* Weapons extension: ma- k++ F+2 X
|
|
|
|
|
In other words, we really haven't reached true AI yet. (I happen to agree with you on this.)
|
|
|
|
|
The big problem: Which human ethics and morals?
The simple answer: Ours! Mine!
The world isn't that simple. Not even "The good guys" (that is: us!) agree on the "right" ethics/morals. It varies greatly from the one end of the political scale to the other. It varies greatly from culture to culture, even today. It varies greatly through the centuries. The only really significant argument in favor of "our" culture/ethics (and I might add: whatever that would be, after we have fought through all the necessary internal wars) is that it is ours.
Probably an unfounded anecdote: Gandhi was once asked his thought of Western civilization, and he responded: That would be a great idea! Even though the story most likely is made up, quite a few people get the message: There are lots of value systems where our Western 'culture' (that includes ehtics/morals) come out on top.
Since we are the masters of digital technology and AI, any attempt to implement ethics/morals into AI systems, is likely to be our Western variant. That is of course OK as a weapon of cultural imperialism, but not necessarily the best for global ethics/moral. Except in our own eyes.
|
|
|
|
|
TL;DR;
As if human ethics and morals were something to be taught...
Addition (to avoid more misunderstandings) copied from my answer below:
Quote: Let me rephrase it:
As if human ethics and moral were good enough to be used as role model to be taught.
M.D.V.
If something has a solution... Why do we have to worry about?. If it has no solution... For what reason do we have to worry about?
Help me to understand what I'm saying, and I'll explain it better to you
Rating helpful answers is nice, but saying thanks can be even nicer.
modified 2-Nov-22 19:47pm.
|
|
|
|
|
When 12 lines (at my current window width) is "TL;", then you put a mark on yourself as someone who cares zero about serious discussions.
And, as with a great number of "TL;DR;" reactions: I am quite sure that you did read all of it, but disliked it. Your "dr;" simply says: I close my eyes for this part; I find it more convenient to ignore it.
Yes: Ethics and morals are something to be taught. You may believe that what you have been taught in those areas are "god given" or "natural", but then, why don't everybody agree with you? They don't. That is not because they have no moral/ethics, but because they have a different moral/ethics. Your ethics/morals is taught, just as much as theirs.
|
|
|
|
|
Yes, I read your whole message.
But you didn't understand mine. I was not criticizing your message. I was criticizing "human ethics and moral" (did you read my following message[^]? I think that would have helped to avoid the missunderstanding)
Let me rephrase the one you have answered:
As if human ethics and moral were good enough to be used as role model to be taught.
M.D.V.
If something has a solution... Why do we have to worry about?. If it has no solution... For what reason do we have to worry about?
Help me to understand what I'm saying, and I'll explain it better to you
Rating helpful answers is nice, but saying thanks can be even nicer.
|
|
|
|
|
obermd wrote: This is why we need to teach our AI systems human ethics and morals. As it would do a difference... do you lately see the news?
M.D.V.
If something has a solution... Why do we have to worry about?. If it has no solution... For what reason do we have to worry about?
Help me to understand what I'm saying, and I'll explain it better to you
Rating helpful answers is nice, but saying thanks can be even nicer.
|
|
|
|
|
obermd wrote: This is why we need to teach our AI systems human ethics and morals. Trump's version or Biden's version or Putin's version or Kim Jong-un version? Heck, even my neighbor and I disagree on ethics and morals.
|
|
|
|
|
I am tired of primitive obsession and the excessive use of primitive types to model a domain. And only, in some cases
|
|
|
|
|
It's a good article. I've certainly suffered pain from using a primitive type when more abstraction was called for.
It links to an older post about "primitive obsession" which uses a zip code as an example, saying that making it a type prevents inappropriate values from being assigned. As it turns out, that's a really bad example. If you have accounts and credit cards in the US but live outside the country, institutions that can't handle anything other than nnnnn (or nnnnn-nnnn) zip codes are a royal buttpain. When I'm in the US and use a US-based credit card at a gas station and have to punch in my zip code, the answer is 00001. There's no such zip code, but it's what one major bank uses for non-residents.
|
|
|
|
|
Google Hangouts is scheduled for death today. The phone app has been individually booting people off the service since July, but the last vestiges of Hangouts, the web app, will be shut down today. I know - stories about Google cancelling things is almost a "sun came up today" kind of story
|
|
|
|
|
In the US everybody has iPhones and use what's there. In Europe everybody has Whatsapp, which is easier as it's linked to phone number, or Telegram, which can be totally anonymous (useful for newsletters or raiding in MMOs, especially before Discord was a thing).
GCS/GE d--(d) s-/+ a C+++ U+++ P-- L+@ E-- W+++ N+ o+ K- w+++ O? M-- V? PS+ PE Y+ PGP t+ 5? X R+++ tv-- b+(+++) DI+++ D++ G e++ h--- r+++ y+++* Weapons extension: ma- k++ F+2 X
|
|
|
|
|
US government looks at ways of limiting the impact of ransomware across the world. To solve: put 1 million USD in unmarked bills behind the third bench in the park
Not pick a park, ALL the parks.
|
|
|
|
|
ZDNet wrote: US government looks at ways of limiting the impact of ransomware across the world. In another words... trying to get the competence out of their business?
M.D.V.
If something has a solution... Why do we have to worry about?. If it has no solution... For what reason do we have to worry about?
Help me to understand what I'm saying, and I'll explain it better to you
Rating helpful answers is nice, but saying thanks can be even nicer.
|
|
|
|
|
As the GitHub Copilot "AI pair programmer" shakes up the software development space, Microsoft's Mads Kristensen reminds folks that Visual Studio's IntelliCode ain't too shabby, either. Not "that one", the other creepy one
|
|
|
|
|
Kind of related?[^]
M.D.V.
If something has a solution... Why do we have to worry about?. If it has no solution... For what reason do we have to worry about?
Help me to understand what I'm saying, and I'll explain it better to you
Rating helpful answers is nice, but saying thanks can be even nicer.
|
|
|
|
|
Very
TTFN - Kent
|
|
|
|
|
Many organizations are struggling with artificial intelligence deployments despite believing that AI will be critical to business success over the next five years, according to a report by Deloitte. AI is neither artificial nor intelligent. Discuss
|
|
|
|
|
Quote: – to 'middling results' What do you expect when you don't use quantum AI?
|
|
|
|
|
David O'Neil wrote: What do you expect when you don't use quantum AI? To save a cat's life?
M.D.V.
If something has a solution... Why do we have to worry about?. If it has no solution... For what reason do we have to worry about?
Help me to understand what I'm saying, and I'll explain it better to you
Rating helpful answers is nice, but saying thanks can be even nicer.
|
|
|
|
|
Only half the time. Possibly.
|
|
|
|
|
NSA, CISA, and the Office of the Director of National Intelligence (ODNI) have shared a new set of suggested practices that software suppliers (vendors) can follow to secure the supply chain. Tip 0: let us in, or we'll huff and puff
|
|
|
|