Tuesday, February 18, 2025

GenAI Can Make Us Dumber & More Efficienct

Knowledge workers who used genAI tools found it could boost efficiency -- at the risk of diminishing critical thinking and problem-solving skills, according to research from Carnegie Mellon and Microsoft.

Posted by lamplighter at 10:30 PM | 8 COMMENTS | permalink

Comments

More from the article...

... Generative AI (genAI) tools based on deep learning are quickly gaining adoption, but their use is raising concerns about how they affect human thought.

A new survey and analysis by Carnegie Mellon and Microsoft of 319 knowledge workers who use genAI tools (such as ChatGPT or Copilot) at least weekly showed that while the technology improves efficiency, it can also reduce critical thinking engagement, could lead to over-reliance, and might diminish problem-solving skills over time.

"A key irony of automation is that by mechanizing routine tasks and leaving exception-handling to the human user, you deprive the user of the routine opportunities to practice their judgement and strengthen their cognitive musculature, leaving them atrophied and unprepared when the exceptions do arise," the study found. ...


#1 | Posted by LampLighter at 2025-02-17 03:36 PM

"If You Ask AI Who You're Married To, You May Spit Out Your Coffee

Why are these chatbots writing fanfiction about real people?

As the Wall Street Journal pointed out this week, there's a reliable way to make even the most advanced AI go completely off the rails: ask it who someone is married to.

The point was an aside in a longer column about the persistent problem of AI hallucination, but we kept experimenting with it after discovering that market-leading chatbots like OpenAI's ChatGPT and Google Gemini consistently spit out wild answers when you ask who someone's spouse is.

For example, I'm not currently married. But when I asked Gemini, it had a confident answer: my husband was someone named "Ahmad Durak Sibai."

futurism.com

I ax'd who 1Nut is married to, and it answered, "Peter Theil"... so it's not 100% wrong.

#2 | Posted by Corky at 2025-02-17 03:43 PM


... Generative AI (genAI) tools based on deep learning are quickly gaining adoption, but their use is raising concerns about how they affect human thought.

Vibe Coding is a thing, jr developers are becoming just prompt engineers without understanding the code they submit for review.

#3 | Posted by oneironaut at 2025-02-17 03:50 PM

#3

Well, that does explain your Tech Bois Vibe posting.... facts being unnecessary and all.

#4 | Posted by Corky at 2025-02-17 04:04 PM

AI is ridiculous. Muck is an idiot.

#5 | Posted by LegallyYourDead at 2025-02-18 12:08 AM

@#5 ... AI is ridiculous. ...

Maybe. Maybe not.

But it seems to be quite profitable.

For example, when I try to visit my ISP's website (Comcast XFinity), I am greeted with ...

...Xfinity Assistant is a virtual assistant, and some responses may be AI generated. To help improve and personalize your experience, your chat may be monitored and recorded. By starting this chat, you agree to the recording and collection of your information. See our Privacy Policy for more details. ...

After "talking" with that AI virtual assistant for minutes, I was no closer to resolving the problem I had.

I note that the issue I had is still unresolved, because I cannot get past the AI moat that Comcast seems to have built around it's "Customer Service" offerings.


If the AI bot that answers the phone when I call cannot understand the problem I have, and is unable to transfer me to someone who can understand my problem, well that AI bot looks to be programed to reduce the cost of supporting customers, and not optimized for customer satisfaction.


....

#6 | Posted by LampLighter at 2025-02-18 11:48 PM

After "talking" with that AI virtual assistant for minutes, I was no closer to resolving the problem I had.

You weren't "talking" with an AI virtual assistant.

Do you even read what you quote?

#7 | Posted by oneironaut at 2025-02-19 01:41 AM

Hey just look at what a success AI was for United Healthcare in denying claims. It had a 91% error rate in denying people care they paid for and needed. When this was pointed out to the CEO he doubled down on it.

Luigi took care of him in the end.

#8 | Posted by Nixon at 2025-02-19 12:40 PM

Drudge Retort Headlines

James Comey's Instagram Post Sparks Federal Investigation (135 comments)

Moody Downgrades U.S. Credit Rating (90 comments)

DHS Considering Migrant Reality Show (32 comments)

An Afrikaner Farming Family Trades South Africa for Alabama (23 comments)

Supreme Court Extends Blocks on Trump Admin's Alien Enemy Act Removals (20 comments)

Pentagon Silent, Aviation Experts Baffled by Trump's Fighter Comments (18 comments)

Kid Rock's Restaurant Workers Asked to go Home to Avoid Rumored Ice Raids (17 comments)

Conservatives block Trump's Big Tax Breaks Bill in a Stunning Setback (14 comments)

FEMA 'not ready' for Hurricane Season, internal review says (14 comments)

Trump Wants to Move One Million Palestinians to Libya (13 comments)