AI /Education
 

IMO It isn't AI that people should be worried about in Education. The biggest impact will be with religion in the public education sector and historical revisionism.

I share doomerists concerns that competing AIs may (will?) be used for nefarious purposes and will be highly disruptive to mankind if not an outright existential risk (note: I only use the word existential to describe what I believe are actual existential risks). Here you have a novel system that runs at full speed making them exponentially more productive. This system may soon outpace and outmaneuver humans, and have powers of persuasion that enable them to bypass safety brakes. Add recursive self-improvement, a stratospheric knowledge base that will allow them to synthesize insights, draw correlations and become master prognosticators, an unearthly evolutionary trajectory, ***unknown emergent behaviors,*** and coordination at scale whether acting solo or with other AIs and you have a looming disaster in the making. It could explain Fermi's Paradox. It could also be quite good...in the short term.

The brightest among us don't understand what really goes on in the 'black box' and that black box is expected to morph into a singularity.

There is no room for error yet here we are, plowing at lightning speed in our quest in who can create the most god-like machine. Most folks have no idea what is down the pike.

AI. You either embrace it, or get left behind.

New hobby :-)
 

Wow. Had no idea. Had merely learned to ignore that first answer from AI on my Google searches, having found too many blatant errors in them. This post of yours is the most interesting thing I've heard - by far. Thanks!

I agree, I find it much easier to go to a golf pro.

Interesting how some people seem worried about the potential of AI. I think that it has the potential for misuse by people such as fusion had. But that AI may develop a super consciousness and try to master its own destiny — naw I don’t see that happening. A long way for a machine built on zeros and ones to go. Until it becomes a living being, and even then, not just any living organism… The idea comes straight out of sci-fi.

Want to defeat AI, just ask it to play tic-tac-toe against itself. That will keep it busy while you sneak around the back to unplug it.

My PCP works for one of the world's greatest hospitals, Mass General in Boston. He's been using AI for at least 2 years. I was there the week before my annual migration to TV and watched him type in question with symptoms, etc ... it spit out a diagnosis in about 5 seconds.

I opted to forgo the personality transplant that was suggested. Too late in life for a do-over.

Let me take that in sections -

1) I wouldn't listen to Gates, he's long been out of it as well as being his own orchestra, and by the way he's not all that technical either. (yes I've spent time with him in Seattle) he is a saleman and a real good one.

2) AI is still new, and being jumped on much like cloud was in IT. The problem is power in 2 ways. AI having more power (not good), and power to run AI (aka Nvidia's new chip).

The thing to watch out for is what we said back in the 80's when AI was tossed around (for those in the business - this is back when Prolog was introduced). Artificial Intelligence will only be as intelligent as the idiot that programmed it. This still holds true today. Let's take Watson (IBM) this was the first entry in an "AI Like" platform and for the most part it failed. Reason I said that was it's primary intention was to use in the Medical arena. It in turn wasn't fed all that accurate information....which leads me to..

3) Using AI for health services. I'd consider this to be quite dangerous in many ways, much like Watson failed, so will AI. The reason is that every human is different which means with every human the 'basis' changes, including the unknown. Do you know if you are not allergic to everything produced by Pharma? Actually you don't, because there are so many different drugs, antibiotics, dose levels, etc out there you could never know 100%. Currently the public expects AI to work at 100%, in IT "it works" is 100% until we want the public to consume something for our own good (cloud). Face is 99.9% is termed "it doesn't work" from a true strict software testing perspective...and the shoe drops - IT doesn't really test software like they used to. They would rather shove it into your hands and let you find the problems - it's less expensive (enter India, Brazil, Philippines) for the company. IT has been doing it for decades. Why does Windows 11 stil have issues? easy answer Microsoft doesn't really test it, you do.

So when you end up in the ER because of AI.... WHO will take that call? India? LOL how ridiculous (as it already is).

Bottom line - don't run to it, crawl very cautiously. Cloud still sucks, costs more than having the equipment in house, and isn't 100% uptime like they promised. So what do you want from AI? You will get the same result. IMHO I don't want to accept mediocre at 'perfect' pricing like the rest of corporate america has sucked into.

That's what almost 50 years of IT does to you - welcome to reality. and no the government isn't even close to keeping anything safe or managed properly they are still "working on it"

The word "embrace" carries a certain connotation that I don't think necessarily applies. For example, one need not "embrace" gun culture to know what guns are or how to use them if/when necessary. (ref: Quigly, down under. "I said I don't like hand guns. I never said I couldn't use one")

At some point it might be useful to distinguish between

lookups (google) - finding information that matches the actual words in the questions
expert systems (many automated help desks) - traversing a pre-programmed decision tree to follow the right branch
AI (chatGpt and others) - learning systems that are trained on large amounts of data and return answers based on that training

Lookups are as good as the user who consumes the returned data. Expert systems are as good as the "expert" who created the decision tree. AI is as good as the training material and the trainer which in some ways is a combination of the previous two.

To me, the most frightening thing about AI is the level of confidence people are likely to give the results. Some very specialized AI will be trained on tightly controlled data sets by true experts in the field. Much AI will be trained on the internet with all its bad information, trained by a human with all their inherent biases, and trained in a hurry in order to make a buck - not a system that should inspire the confidence it will likely be given.

My personal pet peeve is when AI is used as a presenter or narrator for various shows and constantly mispronounce words/names, oftentimes and in a variety of ways. Eg, a word/name like Bob, becomes boa-b, or Paris might become pay rise.

Go off grid….:shocked:

What I keep reading, our children are on a low learning plateau already!

It is surprising that Tom Nichols is held in such regard since he seems liberal when I see him on TV.

No privacy while using AI search engines.
 

Data Linked to You
The following data may be collected and linked to your identity:
Purchases
i Contact Info
• User Content
• Browsing
History
all Usage Data
Other Data
1 Location
@ Contacts
Search
History
Identifiers
Diagnostics