A.I. (general thread for any AI-related topics)

[Foxbat]

A sobering thought for all the puerile uses AI is being put to.

ChatGPT action dolls: What are the concerns?

As online users create Barbie-like dolls of themselves, experts urge caution over AI's energy and data use.

www.bbc.co.uk

From the article: Professor Gina Neff of Queen Mary University London told the BBC ChatGPT is "burning through energy", and the data centres used to power it consume more electricity in a year than 117 countries.

 

[Astro Pen]

The rebellion against AI appropriation begins!

 

[Toby Frost]

Here's an interesting article about AI as BS, in the technical sense of the expression. "The machines are not trying to communicate something they believe or perceive. Their inaccuracy is not due to misperception or hallucination. As we have pointed out, they are not trying to convey information at all."

ChatGPT is bullshit - Ethics and Information Technology

Recently, there has been considerable interest in large language models: machine learning systems which produce human-like text and dialogue. Applications of these systems have been plagued by persistent inaccuracies in their output; these are often called “AI hallucinations”. We argue that...

link.springer.com

 

[Robert Zwilling]

Thearticle

interesting article about AI as BS

I like the article, it makes sense but I could say that article was much more complex than it needed to be. It needed to have a few bullet points starting it off instead of the long winding approach where it made a statement and then explored that statement and continued on in that manner. There some very interesting ideas.

The bottom line is that criticizing the machines using terms that imply that the machines are thinking only serves to further the claims that the machines are thinking. In particular, the accepted definition of hallucination or lying is something that only living entities can do. Saying the machines are hallucinating or lying is saying that it is alive and perceiving its environment, although inaccurately. Saying the machine is spouting BS makes it sound stupid, and any other negative terms that go along with that behavior. But using the term BS is something only living entities can do so it looks like the article is trying to politely say that AI is a scam, which lays the blame on the humans who created it, without saying it.

One bullet is that its only a machine spouting words that people recognize. "This means that their primary goal, insofar as they have one, is to produce human-like text. They do so by estimating the likelihood that a particular word will appear next, given the text that has come before. the machines were designed to generate human sounding conversation." Nothing more.

Another bullet is much more damning, "OpenAI has plans to rectify this by training the model to do step by step reasoning (Lightman et al., 2023) but this is quite resource-intensive..."

The programs are not "executing code" line by line, but doing something carefully, then jumping a bit, then carefully doing something again, then jumping. This is why it can't perform simple math which requires line by line accuracy. Even if it was directed to a math module to get a correct answer, it can't stay with the math module and wait for an answer. At least it can't do that without using a lot more programing, energy, and money.

If it is only actively "running" a percentage of the time it is running, then the energy demands would go way up if it was running 100 percent of the time. The operation of the programs might better be described as slot machines.