Nobody cares about "AI" (Chatbot: I disagree.)

[Free]

All Science journals will now do an AI-powered check for image fraud

It will only catch the most blatant problems, but it's definitely overdue.

arstechnica.com

On Thursday, the research publisher Science announced that all of its journals will begin using commercial software that automates the process of detecting improperly manipulated images. The move comes many years into our awareness that the transition to digital data and publishing has made it comically easy to commit research fraud by altering images.

While the move is a significant first step, it's important to recognize the software's limitations. While it will catch some of the most egregious cases of image manipulation, enterprising fraudsters can easily avoid being caught if they know how the software operates. Which, unfortunately, we feel compelled to describe (and, to be fair, the company that has developed the software does so on its website).

 

[Dakota Tebaldi]

George Takei :verified: 🏳️‍🌈🖖🏽 (@georgetakei@universeodon.com)

Attached: 1 image This feels like a Star Trek episode.

universeodon.com

 

[Essence Lumin]

George Takei :verified: 🏳️‍🌈🖖🏽 (@georgetakei@universeodon.com)

Attached: 1 image This feels like a Star Trek episode.

universeodon.com

I didn't know chatgpt was written in Python. I guess that lays to rest the idea that Python is no good for large programming projects.

 

[Dakota Tebaldi]

I don't think it is. "You are a Python interpreter" is like a roleplay instruction, for lack of a better way to put it. As in, "respond to the following as if you were a Python interpreter".

 

[Essence Lumin]

I don't think it is. "You are a Python interpreter" is like a roleplay instruction, for lack of a better way to put it. As in, "respond to the following as if you were a Python interpreter".

This Quora article mentioning GPT3 says it is. Of course, GPT3 is old now and Quora is not the most reliable of sources.

What’s the code for Chat GPT look like? How many lines?

Answer (1 of 5): This may be long but short answer is, millions.. Let me explain: It’s hard to point at a single ChatGPT codebase, it’s really just a product of an entire ecosystem of libaries, services and tooling. At the very core, the transformer model and training scripts (in PyTorch) are on...

www.quora.com

`The code for GPT-3 is written in Python, and it uses the PyTorch library for its machine learning and deep learning operations. The codebase consists of several different components, including the preprocessing and tokenization of input text, the training of the model, and the generation of text.

 

[Noodles]

Almost all of these AI tools floating around use and require PyTorch when running locally. It would make sense that they are running Python.

Wikipedia says there is a C++ version too though.

https://en.m.wikipedia.org/wiki/PyTorch

 

[Free]

 

[Dakota Tebaldi]

I laughed at this for way too long

 

[Free]

OpenAI claims The New York Times tricked ChatGPT into copying its articles

OpenAI insists training AI models on copyrighted data is fair use.

www.theverge.com

OpenAI has publicly responded to a copyright lawsuit by The New York Times, calling the case “without merit” and saying it still hoped for a partnership with the media outlet.

In a blog post, OpenAI said the Times “is not telling the full story.” It took particular issue with claims that its ChatGPT AI tool reproduced Times stories verbatim, arguing that the Times had manipulated prompts to include regurgitated excerpts of articles. “Even when using such prompts, our models don’t typically behave the way The New York Times insinuates, which suggests they either instructed the model to regurgitate or cherry-picked their examples from many attempts,” OpenAI said.

OpenAI claims it’s attempted to reduce regurgitation from its large language models and that the Times refused to share examples of this reproduction before filing the lawsuit. It said the verbatim examples “appear to be from year-old articles that have proliferated on multiple third-party websites.” The company did admit that it took down a ChatGPT feature, called Browse, that unintentionally reproduced content.

 

[Free]

Uh boy...

YouTube is cracking down on AI-generated true crime deepfakes

The rule appears to respond to AI true crime content.

www.theverge.com

YouTube is updating its cyberbullying and harassment policies and will no longer allow content that “realistically simulates” minors and other victims of crimes narrating their deaths or the violence they experienced.

The update appears to take aim at a genre of content in true crime circles that creates disturbing AI-powered depictions of victims — including children — that then describe the violence against them. Some of the videos use AI-generated, childlike voices to describe gruesome violence that occurred in high-profile cases. Families of victims depicted in the videos have called the content “disgusting.”

YouTube’s policy update will result in a strike that removes the content from a channel and also temporarily limits what a user can do on the platform. A first strike, for example, limits users from uploading videos for a week, among other things. If the policy is violated again within 90 days, penalties increase, with the eventual possibility of having the entire channel removed.

 

[Free]

How Amazon Fashion is using AI to help you find the perfect fit

4 AI-powered innovations—personalized size recommendations, review highlights, re-imagined size charts, and fit insights—are helping customers shop with confidence.

www.aboutamazon.com

 

[Noodles]

How Amazon Fashion is using AI to help you find the perfect fit

4 AI-powered innovations—personalized size recommendations, review highlights, re-imagined size charts, and fit insights—are helping customers shop with confidence.

www.aboutamazon.com

I think this is the thing that is like 50% of my Reddit ads.

 

[Bartholomew Gallacher]

The Austrian ministry of work created a chat bot based on ChatGPT, which should act as job advisor for young people in school.

It exactly went down the drain as most would guess: the creation of that thing did cost around 300.000€. For that project budget the contractor created a chat prompt with exactly 3 (!) rules to follow. So, 100.000€ for every rule, pure genius!

It advised young male to pursue a career in IT, while young women should embrace going into gender studies. And shortly after that being service personnel and cook was advised.

The thing is a commercial off the shelf solution of a company named Goodguys.ai. And it is full of prejudice and other shady stuff, a perfect mirror of the bias in internet.

Give it a try here, just tell it to speak English upfront: Berufsinfomat | AMS

 

[Free]

Artists are making creative companies apologize for using AI

Businesses that rely on creatives should probably avoid angering them.

www.theverge.com

Popular drawing tablet manufacturer Wacom is the latest target being slammed by the digital art community after appearing to use AI-generated images in ads. Over the weekend, creatives across X (formerly Twitter) and TikTok noticed that Wacom was promoting its Intuos pen tablet with a dragon illustration that showed telltale marks of AI-generated images — such as questionable scale designs and fur blending unnaturally into other sections of the image.

Wacom deleted the images without explanation, fueling speculation that an industry-standard brand for artists was using tools widely criticized for replacing them. And it wasn’t the only AI controversy this weekend. Wizards of the Coast (WotC), the publisher behind Magic: The Gathering and Dungeons & Dragons, also issued an apology on Sunday for using an ad with AI-generated elements. The controversies have escalated mistrust around an already complicated question: how can creatives, and the companies that work with them, navigate a flood of images that are easy to create and hard to conclusively detect?

 

[Noodles]

The thing is a commercial off the shelf solution of a company named Goodguys.ai. And it is full of prejudice and other shady stuff, a perfect mirror of the bias in internet.

Give it a try here, just tell it to speak English upfront: Berufsinfomat | AMS

This is just a natural problem with all these LLMs and "AI" in general.

It's not actually intelligent.

It just regurgitates what it has been told. And in this case, if you say, give it a bunch of education and job statistics, it's going to say, "Men seem to go into IT a lot and women seem to become cooks a lot."

They just amplify existing bias, because the human created past, was created by bias.

 

[Kamilah Hauptmann]

 

[Cristiano]

Here’s what we know about the SAG-AFTRA AI voice acting licensing deal

AI voice acting coming soon to a video game near you.

www.theverge.com

 

[Dakota Tebaldi]

OpenAI claims The New York Times tricked ChatGPT into copying its articles

OpenAI insists training AI models on copyrighted data is fair use.

www.theverge.com

I see OpenAI is going the Twitter route. "They had to manipulate the algorithm into serving that result!"

They seem to think it's disqualifying that the NYT asked ChatGPT to regurgitate the articles. But the whole problem is, it shouldn't be "regurgitating" them at all, even when asked to.

The Times' case really attacks one of the biggest rote defenses of generative AI - that it doesn't copy and reproduce training material in a cut-and-paste way but rather is "inspired by" training material the way an author or artist is "inspired by" existing works and then makes an original by borrowing styles and elements and ideas. But like, increasingly that's shown to not be true. The Times' examples kind of prove beyond any doubt that the whole line for line originals are "in there somewhere" and will be returned with the right prompt.

 

[Free]

Why?

AI-Generated George Carlin Drops Comedy Special That Daughter Speaks Out Against: ‘No Machine Will Ever Replace His Genius’

Stand-up comedian George Carlin has been brought back to life in an artificial intelligence-generated special called 'I'm Glad I'm Dead.'

variety.com

More than 15 years after his death, stand-up comedian George Carlin has been brought back to life in an artificial intelligence-generated special called “George Carlin: I’m Glad I’m Dead.”

The hour-long special, which dropped on Tuesday, comes from Dudesy, a comedy AI that hosts a podcast and YouTube show with “Mad TV” alum Will Sasso and podcaster Chad Kultgen.

“I just want to let you know very clearly that what you’re about to hear is not George Carlin. It’s my impersonation of George Carlin that I developed in the exact same way a human impressionist would,” Dudesy said at the beginning of the special. “I listened to all of George Carlin’s material and did my best to imitate his voice, cadence and attitude as well as the subject matter I think would have interested him today. So think of it like Andy Kaufman impersonating Elvis or like Will Ferrell impersonating George W. Bush.”

Uh. Whatever...

 

[Innula Zenovka]