Bad uses / over-promised of AI

Forums › Artificial Intelligence (AI, LLM, ML) › Bad uses / over-promised of AI

1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11

55 posts

Master Geek
+1 received by user: 59

ID Verified

#3437635 26-Nov-2025 11:28

This article argues LLMs are ill-suited to AGI and attempts to explain why by drawing on a scientific paper that argues intelligence is cognitively distinct from language use, i.e. thinking and language are not the same.

The article is not a deep dive. Nor is it a review of all the main debates about the relationship between language and thought. The article will be one of many in the next few years, I suspect.

ANglEAUT

altered-ego
2436 posts

Uber Geek
+1 received by user: 842

Trusted

Lifetime subscriber

#3437665 26-Nov-2025 13:20

Copilot in Excel starting to recommend a solution to a formula throwing an error:

Please keep this GZ community vibrant by contributing in a constructive & respectful manner.

davidcole

6099 posts

Uber Geek
+1 received by user: 1465

Trusted

#3437674 26-Nov-2025 13:39

Gave siri a bunch of text to summarise......got told "you have no messages"

Previously known as psycik

Home Assistant: Gigabyte AMD A8 Brix, Home Assistant with Aeotech ZWave Controller, Raspberry PI, Wemos D1 Mini, Zwave, Shelly Humidity and Temperature sensors
Media:Chromecast v2, ATV4 4k, ATV4, HDHomeRun Dual
Server Host Plex Server 3x3TB, 4x4TB using MergerFS, Samsung 850 evo 512 GB SSD, Proxmox Server with 1xW10, 2xUbuntu 22.04 LTS, Backblaze Backups, usenetprime.com fastmail.com Sharesies Trakt.TV Sharesight

gzt

18710 posts

Uber Geek
+1 received by user: 7841

Lifetime subscriber

#3437774 26-Nov-2025 18:02

National Party fake ai advertising - pretty close to impersonating police officers imo

https://newsroom.co.nz/2025/11/25/national-deletes-ai-video-of-[fake]-police-seizing-gang-patches/

Just stupid.

Dynamic

4017 posts

Uber Geek
+1 received by user: 1853

ID Verified

Trusted

Lifetime subscriber

#3438518 29-Nov-2025 14:25

I took a photo of a receipt intending to upload it to Xero but wanted to crop it first and reduce it's resolution. With Microsoft showing AI down our throat, I decided to try the Copilot feature in the Photos app built into Windows 11, and gave the following instruction:

This is a photo of a receipt from a retail store. Can you please crop the photo so only the receipt shows?

The result was very good. I was about to upload it to Xero and then spotted that the top product line did not look familiar. None of the other lines looked familiar either.

The image you gave me to download is NOT my receipt. I don't understand what has gone wrong with my request?!?!

Thanks for pointing that out, Mike — you're absolutely right to be confused. What happened is that I mistakenly generated a new image of a receipt instead of cropping the actual photo you uploaded. That’s not what you asked for, and I appreciate your patience.
Let me fix this properly by cropping your original photo so it shows only the receipt. Hang tight…

I wait. And wait. And wait. Nothing happens.

That's enough AI for today. 😒

“Don't believe anything you read on the net. Except this. Well, including this, I suppose.” Douglas Adams

Behodar

11107 posts

Uber Geek
+1 received by user: 6100

Trusted

Lifetime subscriber

#3438521 29-Nov-2025 14:32

I got annoyed yesterday. I'd done a Google search for how to accomplish something in a piece of software, and its AI popped up and gave me some steps that didn't work, referencing a file that didn't exist. It turns out that it had all changed in version 6, a number of years ago (we're now on version 10).

When trying to find a real search result for it, I found someone who asked exactly what I wanted just a couple of weeks ago. Someone replies with a passive-aggressive "You could just ask ChatGPT instead of posting on a forum" and proceeds to paste ChatGPT's "solution"... which references the same steps that don't work in version 6 or newer.

HP

Shop now for HP laptops and other devices (affiliate link).

gzt

18710 posts

Uber Geek
+1 received by user: 7841

Lifetime subscriber

#3447528 27-Dec-2025 17:06

Several NZME newspapers print recommended Christmas list of non-existent books:

https://www.1news.co.nz/2025/12/27/nzme-apologises-for-christmas-list-featuring-made-up-book-titles/

gmball

578 posts

Ultimate Geek
+1 received by user: 302

#3447530 27-Dec-2025 17:17

Auckland parks promoting the environment and caring for regional parks - surely they have some actual photos of Long Bay Regional Park they could substitute in place of AI garbage?

gzt

18710 posts

Uber Geek
+1 received by user: 7841

Lifetime subscriber

#3447531 27-Dec-2025 17:25

The picture is ethnically homogeneous when compared to the actual crowds at Long blBay. Also no picture of Long Bay is complete without the continuous traffic jam and heat lines.

RunningMan

9193 posts

Uber Geek
+1 received by user: 4845

#3447532 27-Dec-2025 17:42

Pity the AI didn't know how many fingers to put on a hand.

Behodar

11107 posts

Uber Geek
+1 received by user: 6100

Trusted

Lifetime subscriber

#3447534 27-Dec-2025 18:04

I was actually looking to buy a T-shirt the other day, but incredibly they had no photos of it, just AI slop of what it might look like. Unsurprisingly I left the site without buying anything.

Samsung

Shop now on Samsung phones, tablets, TVs and more (affiliate link).

Behodar

11107 posts

Uber Geek
+1 received by user: 6100

Trusted

Lifetime subscriber

#3452801 14-Jan-2026 10:47

A few years ago if I were to accidentally type ".newt" instead of ".net" on the end of a URL then it would tell me that the URL isn't valid. Now it pops up an "AI" summary about environmental monitoring involving newts. This is supposedly progress.

eracode

Smpl Mnmlst
9336 posts

Uber Geek
+1 received by user: 6215

ID Verified

Trusted

Lifetime subscriber

#3452805 14-Jan-2026 11:13

RunningMan:

Pity the AI didn't know how many fingers to put on a hand.

The AI did it deliberately to support diversity and, in particular, people with disabilities. It’s unkind of you to publicly point out that guy’s hand issue.

Sometimes I just sit and think. Other times I just sit.

Tinkerisk

4809 posts

Uber Geek
+1 received by user: 3669

#3452808 14-Jan-2026 11:22

Hackers attack with the help of AI.
Admins defend with the help of AI.
Admins must arm themselves against as many attacks as possible (impossibly all of them).
Hackers only need to find a single point of attack.
This race goes on forever.

Therefore, one should behave as if there is always someone on the LAN (WiFi!) reading along, even if this is not actually the case, and there is no need to be paranoid.

Qui nihil scit, omnia credere debet. - He who knows nothing must believe everything.
Firewalls do NOT stop dragons.
Avoid Big Tech!
In effect we have everything to hide from someone, and no idea who someone is.

kingdragonfly

11996 posts

Uber Geek
+1 received by user: 12895

Subscriber

#3452869 14-Jan-2026 12:35

"Don't Call It Slop!" - The Corporate AI Dumpster Fire

Upper Echelon

1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11