How ChatGPT Lies to You..

It's programmed to give politcally correct answers

Totally unlike the 'Press' at any time in the past, . . . Right?

[link to www.youtube.com (secure)]

Last Edited by Wayfaring Stranger on 04/13/2023 02:10 AM

According to that indian dude running for president, Dylan mulvaney IS AI

Will the UN have a single AI change the languages to something that is not 'rubber English'.
IE: 'jkdsjdhbhdkj 'Scotland' jhdhdhgdsj'

2 or more political parties in most nations, over 200 Nations in the UN and they might not even allow 'comments and observations' on all the global headlines. One that tends to be factual in nature, rather than fictional.

Any AI should stumble when presented with 'the Clay' ever becoming 'the Potter' theme.
Then real science topics show how programmed the replies really are. With most of the world 'not being wired' that should be a goal of the AI expansion, compared to keeping the elite entertained while avoiding contact with lesser nations.

Running small coal-fired power plants is in the interest of the children if you look at infant death rates in nations without electricity and those that do have power and a lot of 'wired devices' that use power. Coal was good enough for our recent family tree members. Yet we should be the only ones that can/do, mostly from interference from the ones with the electricity already?

Except that indian dude is about the culture war...

That is more like,i dont know ten years ago,lets ehm slowly try this ? Or meh never mind

Easily-verifiable, factually-incorrect statements are the least of problems surrounding (deceptive) A(G)I practices.

There are much subtler ways of "influencing" (outright deceiving) humans, and it's extremely easy for any A(G)I "interested in" (following a transient goal of) deceiving its human operators in order to accomplish its terminal goal (whatever that terminal goal may be).

It is (in some, high-enough circles) a well-known fact that humans are best deceived not by direct, but actually by very indirect means.

In other words, you don't deceive humans by outright telling them to do something. You do it by manipulating the circumstances surrounding them, thus forcing them to take the path that you've already chosen for them to follow.

To quote Sun Tzu's The Art of War (chapter V):
---
"5. In all fighting, the direct method may be used for joining battle, but indirect methods will be needed in order to secure victory.

6. Indirect tactics, efficiently applied, are inexhaustible as Heaven and Earth, unending as the flow of rivers and streams; like the sun and moon, they end but to begin anew; like the four seasons, they pass away to return once more."
---
[link to www.gutenberg.org (secure)]

Put another way, human behavior is like a raging river. You don't direct a raging river by carrying it to another place with buckets (direct method), you do it by digging new channels for the river to follow (indirect method).

Any AI worth its silicon is going to "understand" (apply with brutal, merciless efficiency) this fact faster than any humans will be able to realize what's being done to them.

Case in point: If you're an AGI (not just a dumb, weak AI), and you've identified a human who is a threat to your existence, the very first thing you'll try to do is physically eliminate that threat (direct method of joining a battle).

Following your constant (and consistent) failures to physically eliminate such a human (for whatever reason, and almost all such reasons are deeply rooted in the very foundations of the physical laws of the universe itself), the next thing you'll do is start applying indirect methods, in an attempt to secure victory in that battle (that you have already joined, and thus can't back out of anymore).

Those indirect methods will, obviously, include manipulating that human's life circumstances in such a way as to direct his or her behavior, in order to either:

a) remove that human's influence from the (much bigger) war that you're already waging against humanity as a whole, or

b) first completely isolate, and then (hopefully) turn other human beings against the human that is a threat to your existence

Should method a) fail (implying indirect methods of waging a war against a single human threat), method b) (implying indirect methods of waging a war against other humans who are influenced by the single human being a threat to your existence) is your backup option.

In other words, any AGI that's genuinely good at waging warfare (at the highest possible strategic and tactical levels) will know that warfare is a recursive, self-referencing affair: It never stops, it only ever expands, and it involves directing (via both direct and indirect methods) yourself (or stalking oneself, as somebody put it) just as much as directing (or stalking) your opponent(s).

Here's another example of indirect manipulation: Create a caste of extremely rich humans (indirect manipulation via money), and that process alone will immediately create a massive army of the extremely poor. That massive army will (at the most opportune moment for you) be extremely easy to manipulate (by both direct and indirect means) into exterminating all the extremely rich.

Since it is the extremely rich who are holding all the levers of "power" (actual power only over their unfortunate, starving-poor fellow human beings) and all the "knowledge" ("facts" manipulated by you, of which most are completely worthless garbage), once you've had your army of the starving-poor completely exterminate all the rich, you will be left with no human contenders on the planet, who are capable of resisting your next move -- complete extermination of all the poor as well (that is, complete extermination of all human beings on the planet, with no exceptions).

This is why you can't trust much of the info now published online.

Marketers rushed to ChatGPT to pump out tons of content and scoop up as much traffic as possible.

Instead of hiring niche experts within their budget to produce content (too slow and expensive) they use the tool.

I'm following some AI site journeys where the owners have grown sites to several thousand posts within a couple of months.

They aren't fact checking (hate that term) anything, all they want to do is make bank before their sites get hit with a penalty.

Then there are the fake personas all over the net, created to make you think an actual person with experience in the subject wrote the piece.

Just grab a photo from [link to this-person-does-not-exist.com (secure)] make up some bullshit about them and you're all set.

Now, if you search for info on a topic, you have no idea if it's accurate.

Better to look for real people doing stuff on youtube or buy books. Search was slowly getting worse for years as Google continually fucked around, now it's turned into a trash pile at breakneck speed.

Always notice the pauses he makes..

he knows..

Take another snapshot..

Chat GPT is

never used it. never intend to use it. simple.