I thought AI-generated art is not copyrightable. If so, any AI-generated art is immediately in the public domain.

It is not copyrightable. DMCA however will still allow you to strike it and passes the buck to you to deal with it in court.

Dec 5, 2023 18:40:39 GMT Gyro LC said:

I thought AI-generated art is not copyrightable. If so, any AI-generated art is immediately in the public domain.

I have a contact in a creative agency, mid journey art can be exclusively licensed

In the US it cannot. Courts ruled that creations not of human creation cannot be copyrighted, and AI art is created by AI not a human. They use prior rulings for animal art copyright disputes.

US Copyright Office also states that for a copyright to be granted the creator needs to be human and AI generated art, being generated by the computer, does not meet the definition of human creation.

People can try to license their AI art, or claim copyright, but courts will not support them in copyright infringement cases.  Why Hollywood was so willing to cede AI stuff in negotiations.  Anything they make with AI, also is public domain.

What are the laws if you own the AI? And in addition the AI isn't freely or commercially available and is used purely by the owner to generate artwork.

Think I found the best use of AI yet. Kit uses several versions of AI as part of a conceptual maze he created for those phone scammers. He programs these things live on stream when scammers are not biting. The all this works is kit collects scam numbers from the community and calls the numbers with a voice changer pretending to fall for the scam. He strings them along, then offers to pay several thousand bitcoins through a weird wallet. He gives them what seems like real information and it sends scammers into the maze, that is an endless set of CAPTCHA and verifications. Once the scammers verify the information they get an error screen that starts it all over. Now he added the AI it transfers them to the AI after layers of CAPTCHA. crazy Now he added the insane AI things just get weirder and weirder.

As for why these guys are putting up with all this, they believe they conned a rich American out of 20+ million dollars.



The gauntlet video that the AI got merged into.



Anyone interested he streams during the American day on twitch at.  Great mix of programming stuff and hilarious improv fuckery with the scammers.  He now has all sorts of fun toys to use like a fake Bank of America account, insane startup stuff like Cornhub, a video site to stream your corn, and a fake crypto Updog.   He plays about a dozen characters with their own schemes and personalities using the voice changer.  Goal of the channel is to tie up scammers for as long as possible so they are not scamming others during that time.  

.

In an absolutely astounding development, Google was forced to switch off their new image generating AI called Gemini.
Why? It refused to generate pictures of white people.
If you asked for a picture of the founding fathers, you would get minorities and women, it is so unbelievable that it is almost satire.

Boo! Down with The Patriarchy!

Wow holy misunderstanding what happened. Gemini has a pause in generating images of people because the AI did not understand racism, and saw no reason that nazis or a pope should not be black. It is offline while they manually override stuff to teach it about race and hard program racism into it.

www.cnn.com/2024/02/22/tech/google-gemini-ai-image-generator/index.html

It actually refused to produce an image of a white family, but would produce any other colour or ethnicity.

Tools are not racist they are only tools.
If people are going to edit the cultural impact of the truth before releasing it, then that is not going to advance society.

Again, did not add why. Gemini was not trained to generate people of any specific race because they did not want it used for hateful content. So that feature was simply not in it. When you asked for people in a setting, it would pick whatever race it wanted.

But since people freaked out, now they will train it for race, and that minorities should not be displayed for many prompts. I assume they do it for women as well while they are at it. Basically teach it sexism and racism to align with how people view the world.  Then it appears this version will be a separate model for only certain countries.

I tried it, you could generate black people, just not white people.

Google " 1943 German Soldier " and tell me what colour the skin is.

There's a certain irony in Google being the most used search engine in the world and the AI that feeds off it, can get it so wrong accidentally. Or is it?

I ran that prompt through the most basic of open source image generators and guess what I got... a 1943 German Soldier of obvious Germanic descent. No Nazis or ugly propaganda, just a white dude in uniform.

If the incredibly move advanced, more precise, more detailed and more intelligent AI is throwing out results obviously well off the mark, I would suggest it ain't necessary the AI itself, but the set of rules it has been given to play within.

This is two issues. For not generating white race stuff when prompted the reply when asked is:

"While I understand your request, I'm unable to generate images that specify ethnicity or race. It's against my guidelines to create content that could be discriminatory or promote harmful stereotypes."

Google does not want their AI to be used for racist reasons.

For the weird race stuff, you need to train AI to understand historic racism and Google simply did not. For it to know a 1943 German soldier needed to be white male, someone would have to teach it that by rating dozens of generated images related to the prompt. AI is not a magic box, it is trained by people.

AI current thoughts.
Shocking to see Tyler Perry pause growing his empire due to AI.
I find it an interesting ethical conundrum to dress an apartment with AI
www.vice.com/en/article/xgwpmk/ai-generated-furniture-real-estate-listings
I also find the notion you could combine www.dignifai.net/https://www.dignifai.net/ and something like Apple Vision Pro to make your world how you want to see it. Religious, conservative? Cover up people's skin or tattoos with how you want to see them. No more letting people expose themselves to you. No shoulders or ankles in my time! Or, reverse it, make it so you see everyone what AI determines they would look like naked. Racist? Ha! AI makes everyone white for you. It could get weird.

c replace " historic racism " with historical accuracy.

Because that is what it is. Despite your select choice of wording.

It's not racism.

It's not racist if a person is portrayed as their actual ethnicity.

A Viking is white. An African tribesmen is black. A Samurai is Japanese. An Indian is India.

Google have obviously made a choice here, and it has backfired.

Stop acting like everyone else is crazy and putting ideology before reality.

Google is always getting exposed for the meme it is.

I still am baffled as to why people used a service that had a boilerplate of promoting diversity, getting shocked it promotes diversity. And like the use cases are insane. If wanted to see historically accurate stuff, why ask an AI instead off just seek out historically accurate stuff?  

As for the racism debate, to train the AI that a viking can only be white, you need someone to rate the picture and only rate the white pictures as accurate and tell it all other imagines of vikings are wrong. This is discriminate learning and the very foundation of what systemic racism is.  And then it would not even be historically accurate as races had mixed by the time of the Vikings.  Black nazis existed early on and helped execute the Jews, white people existed in India throughout written history, and there were known Prussian and European samurai as well.   The reality people claim is the reality of a racist fantasy not historic reality.  

Google confirmed now this was a hallucination. They wanted to promote a range of expression, and as I suspected and said here, the AI stopped producing single race pictures as it learned over time to promote diversity over single race cases, to the point of closing off single race prompts. Google will be clearing up the hallucinations by training the model that some outputs have to be confined to a single race, which is teaching it racism, and adjusting overall tuning parameters.

Feb 23, 2024 22:11:13 GMT @ness said:

Google is always getting exposed for the meme it is.

This is the thing.

An Encyclopaedia is not meant to serve diversity, it is designed to report facts, if your product is providing information, then manipulating algorithms to match your politics is flat out wrong.

I am pretty sure Gemini was never promoted as an encyclopaedia.

I'm a computer guy, I'm tech-savvy, I like geeky stuff.

I don't like AI, or what it's becoming, and I'm not talking about a Skynet scenario which is far closer to fiction than fact. In my view all these AI text generators and image generators are causing more problems than they solve.

Feb 23, 2024 22:34:45 GMT c said:

I still am baffled as to why people used a service that had a boilerplate of promoting diversity, getting shocked it promotes diversity. And like the use cases are insane. If wanted to see historically accurate stuff, why ask an AI instead off just seek out historically accurate stuff?  

As for the racism debate, to train the AI that a viking can only be white, you need someone to rate the picture and only rate the white pictures as accurate and tell it all other imagines of vikings are wrong. This is discriminate learning and the very foundation of what systemic racism is.  And then it would not even be historically accurate as races had mixed by the time of the Vikings.  Black nazis existed early on and helped execute the Jews, white people existed in India throughout written history, and there were known Prussian and European samurai as well.   The reality people claim is the reality of a racist fantasy not historic reality.  

Find me a Black Viking... I'll wait.

Maybe, you found one... maybe. Or evidence there was one. Maybe a stray got picked up somewhere off the tip of Africa and ran with The Vikings. Maybe an indigenous folk from Newfoundland jumped on a ship and were Vikings. But when the historical reality is Vikings originated from Scandinavia.. the reality is, they were white.

I know reality is difficult for you.

It ain't discrimination or systemic racism.

It's reality.

Historical accuracy.

Reality. Despite your best efforts to suggest otherwise.

This isn't discriminate learning and it's not the very foundation of what systemic racism is.

It was taught to distort reality.

Jone's A History of the Vikings notes that a small number of vikings had black skin.  They were blamann who converted over.  The historical reality you would want to make impossible for the fantasy.  This is racism, a believe that people cannot be things simply because of their race.  

If you want a single name, Geirmund Heljarskinn.

Feb 23, 2024 23:41:57 GMT Emperor said:

I'm a computer guy, I'm tech-savvy, I like geeky stuff.

I don't like AI, or what it's becoming, and I'm not talking about a Skynet scenario which is far closer to fiction than fact. In my view all these AI text generators and image generators are causing more problems than they solve.

Tools aren't the problem, they just fall victim to the tendencies of people... And it's people that are the problem.

I know not everyone here pays attention to the stock market, but Tesla is out and the new darling of the tech scene is NVIDIA, the amount of profit they are making through GPU for the AI sector is mind boggling, it shows how much longer the AI explosion has to run.

I'm seeing a lot of talk lately that I haven't taken the time to research that future AI models might move to general computing meaning they are better suited to CPUs than GPUs.

I love the technology space and is shocking to see Intel fall so far technologically, ARM is improving in leaps and Intel is still on the baby steps, if it wasn't for mobile and long term enterprise customers, Intel would be in big trouble

Feb 23, 2024 23:54:58 GMT c said:

Jone's A History of the Vikings notes that a small number of vikings had black skin.  They were blamann who converted over.  The historical reality you would want to make impossible for the fantasy.  This is racism, a believe that people cannot be things simply because of their race.  

If you want a single name, Geirmund Heljarskinn.

Typical. Focusing on a 1% argument when your 99% wrong.

Google fucked up. Move on. If it were a conservative company white washing history, you'd have 30 posts copy and pasted from reddit in readiness calling for the heads of the company.

Isn't this what you are focusing on right now? Generating images of people in non-diverse is a problem that affects almost no users of Gemini. Most did not realize it was even happening until someone made it a woke thing and Murdoch pushed it.

Literally this entire conversation is a 1% issue.

And it was you who brought up there were no black vikings, and show me a black viking, not me.

Like I said it was a hallucination, a problem of all LLM that occurs in the beta stages, which they warn about in the FAQ section.  They are fixing too as I said.

I'm not enabling you further.

Feb 24, 2024 13:17:12 GMT iNCY said:

Tools aren't the problem, they just fall victim to the tendencies of people... And it's people that are the problem.

Yeah, that's what I was getting at. People tend to ruin everything :lol:

I got to hear JR sing My Way, all AI bad mouthing will CEASE AND DESIST.