What is your Opinions on new AI Technology

I detest this rush to AI. It does not produce unbiased results. It still is biased, intentionally or not, by the programmer writing the code.

In the meantime the increased use of AI is eliminating front-line workers' jobs from professions like customer service. At least McDonalds' has tested their AI at the drive-through and abandoned it because of the problems with speech recognition. People apparently were getting handfuls of butter and bacon strips in their ice cream when placing certain orders. The other day I typed in asking where I could buy a Hobie kayak near Laughlin NV on the Brave browser and was presented with AI answering what an inflatable paddleboard is.

I just don't trust it.

 

Interesting about the face-recognizer's ability to correlate faces with lefty/righty politics. I suppose what it really suggests is that leftiness/rightiness tends to be passed down in families (through upbringing, not all by genetics) and so can end up noticeably correlating with things like facial features that pass down in families by genes.

I'm pretty sure the uses that skeer me most at the moment are things like voice cloning and video faking. Seems like we already have a crisis of people falling over themselves to believe stuff they've been given no remotely adequate reason to believe, and if now they'll even be able to "see" a "video" that "shows" it, or "hear" it in the named person's "voice", that won't be getting better.

 

Like any tool or technology, It's like a double edge sword that can cut both ways.

It can be used for both good and bad. It all depends on the user.

 

I'm a little puzzled by the term "artificial". Maybe "synthetic" or "manufactured" would be more succinct? "Artificial" infers it's not real, fake, whatever?

 

Germans have a term for it. "Ersatz", which means replacement or substitute.

Americans incorrectly translate it as "fake."

The English translate the word for China as the Middle Kingdom, when a more appropriate translation is The Central Kingdom.

The Sendero Luminoso was a guerilla group in Peru. It was translate as the Shining Path. A closer translation would be the Enlightened Path.

 

Yeah, I'm not sure 'artificial' is too far off the mark, really. Webster's 3rd gives it about six column inches, and unreality/fakeness don't really figure in to most of the meanings.

Meaning 1 is "contrived through human art or effort and not by natural causes detached from human agency". That might fit pretty well.

Two inches of reading before you get to meaning 4 (a "not genuinely or spontaneously felt or experienced ...", b "AFFECTED, SHALLOW, CONVENTIONALIZED, STILTED: not natural, spontaneous, or free", c "IMITATION, SHAM").

Most of meaning 4 seems geared to describing ways certain people act or present themselves, though (c) carries it on toward general shamminess.

Maybe meaning 4 gets overemphasized these days, at the expense of remembering other stuff the word also means.

The root 'artifice' literally comes from ars and facere, "skill" and "to make", using a skill to make something. It also has a meaning 2 that moves toward trickery, guile, insincerity. And again, it might be that meaning that gets the most airplay these days.

I'm not sure ersatz is a better fit. My Cassell's does translate it as replacement or substitute, but I don't think most people working in AI are really out to have it replace or substitute for our own intelligence (whether or not it ends up doing so. ) I'd say 'artificial' really better captures the key defining point, that it was something made by human skill and effort, rather than arising some other way.

 

I agree with Georgina. Like any new technology, most will use it for good, and some will use it for evil

 

Pretty soon AI will morph into a compound word, the origins forgotten. Say AyeAye??? Or in Canada EhEye...

 

I had fun with some image generators and more recently some silly comedy/country songs, but it looks like a real disaster in the making for information literacy, privacy and socially-important legal standards.

In other words I think there's some neat arts & entertainment hacking going on, but I see danger when this technology is applied to anything that wasn't supposed to be light-hearted fun.

Legal tools like 'photographic evidence' and 'video testimony' may be rendered useless pretty quickly, and things will be quite messy during the transition.

I'm also a bit unnerved by its use in chemistry and materials engineering. I understand some AIs are doing well to solve puzzles of producing novel substances, finding cheaper and easier processes for production.

I see where that could have some very important benefits for industrial and medical applications, and some particularly bad opportunities for weapons development.

 

There are several AI variants. The ones that are Google on steroids without attribution, not so interesting and subject to copyright suits. Nor am I impressed with AI enhanced CGI used to lie. But I am a big fan when used in Full Self Driving.

For the past three months, V12 Full Self Driving has out performed anything else. There are some reproducible edge cases but they are easily identified and handled. But earlier problems that demanded so much ‘defensive’ management have gone down a lot.

Bob Wilson

 

In recent months, manuscript reviews done on behalf of journals have started to have reviewers say how they use AL Which I would prefer to call Computer Intelligence. Uses include detecting plagiarism, seeking information that should have been included in manuscripts to generating text for the reviews themselves.

So journals ask reviewers "what AI did you use and how?". I do not use this. Not a fan. My position is that if I hold myself competent to perform journal review, I do it with my mental resources, not something from some computer programming kid. Dang whippersnappers ...

At the same time I have a need to find 'all' publications on particular topics to create reviews. You would not believe how many 'all' can be. Sometimes titles, keywords and even abstracts don't reveal all that is contained within. So I am perfectly willing to use web crawlers that use text strings I suggest for full-text searching, and let AI try to improve those. There is always wetware (me or collaborators) making decisions about what gets included. Such procedures (all procedures really) get included in manuscript's 'Methods Section'. If anything is not included there, reviewer holds up a yellow card.

==
Summary, I agree with other perspectives given above. AI tools can be used or misused.

 

Exciting times we are living in. especially for those in the tech industries that are keeping up on the latest developments in AI and Quantum.
Not so much for simple users who notice the newer apps doing new things and how to figure out how to get an app to function like they've ( the users ) become accustomed too seeing the apps behave. Even on older phones that don't have the new AI acceleration built into them yet, like the newer models on the market now.
OpenAI / ChatGPT 4 and how it's straddling the open and proprietary licenses, Nvidia new $25 million chips / modules, Google, IBM, Adobe, US, China, Russia battling for the Quantum / AI lead.
Data Center growth projections not only in compute capability also in millions of sq ft per installation. And the amount of energy needed to power and keep at a working temp the electronics.
Current Quantum machines needing to be kept at near absolute 0 to function at all.

Interesting times ! ?

I wonder what it would be like to live in a time frame when ( as it's recorded in the history books ) most every human alive thought the world was flat.

And will humans even make it to Type 1 on the Kardashev Scale?
Kardashev scale - Wikipedia

 

Imagine bullets that look for human blood temperature and steer towards it. After the smart weapons come the brilliant weapons.

 

So far (fortunately?), the stuff that writes text seems to be at the level of writing content that resembles a certain style but is drivel when you try to really read it. Like this blog post I stumbled on, and pretty much have to believe came from an AI prompt:

https://www.bmwofreading.com/spark-plugs-and-how-they-affect-your-bmw/

Anybody who reads all the way through it will probably be set back months or years from learning how an engine really works....

So for the time being, anyway, a lot of AI-generated text contents seems pretty easy to spot, just by exercising our human ability to spot not-quite-really-coherent drivel.

Trouble is, that human ability seems on the decline, too. There's a convergence in progress toward incoherent drivel. AI's capabilities are rising toward it, and humans' expectations are falling toward it.

 

"Exciting times we are living in. especially for those in the tech industries ..."

This is a narrow view. Food, water, energy and medical (FWEM) are incompletely and very unevenly made available. Even though we pretty much know that global economy would benefit/increase if everyone could have 'their hands in'.

Eight billions will become 10 soon enough, while climates move away from what we are used to. FWEM will be affected, not entirely favorably.

Meanwhile most of the Richest look to how they got that way, rather than towards a better future.

Meanwhile Wars for No Good Reason persist. Because they used to seem beneficial, I guess.

No Kardashev level ANYTHING civilization could cling so obsessively to its past.

==
Silicon-enabled computer technology offers us help. No doubt. But it does not exist independently from all the important other things also happening.

 

you mean we'll finally be rid of mosquitos?

I was against AI but now I'm on the fence.

with that joke out of the way, I think I'm reading that a number of us share the idea that AI can empower and amplify lots of things. It just depends on whether you're trying to accomplish good clean fun or >>SUBJUGATE ALL HUMANS<<

 

I see them used for land mine detection and removal. But the Ukrainians also have remote, flail, mine removal vehicles.

Bob Wilson

 

I agree:

If the text AI systems included footnotes for every phrase they rip, no problem. But automated plagiarism is never OK . . . having been plagiarized before.

Bob Wilson

 

And the subjugate part isn't coming from the machines.