This New York Times story is behind a paywall, but for those who can read it, I think it shows a clear parallel between concerns of photographers, illustrators, videographers, musicians about AI tools building a business on the backs of other people's work without compensating them.

This story is about writers for movies and TV: "Will a Chatbot Write the Next �Succession�? As labor contract negotiations heat up in Hollywood, unions representing writers and actors seek limits on artificial intelligence." 98% voted to approve a strike.

https://www.nytimes.com/2023/04/29/business/media/writers-guild-hollywood-ai-chatgpt.html

One significant difference is that writers have a union representing them.  There are various points of leverage mentioned in the story which don't apply to stock agency contributors. One is the ruling that you can't copyright something generated by AI (referring to visual art) which may help keep human writers employed as the lack of copyright protection would discourage large financial investments in a project - anyone could then just copy it and the company with the investment would have no recourse.

There's also the writer's credit: "...it falls to the Writers Guild and not the studio to determine who receives a writer�s credit on a project, and that the union will guard this rite jealously. �We want to make sure that an A.I. is never one of those writers in the chain of title for a project,� "

So far, it seems that musicians and TV writers have large organizations who will go to bat for creator's rights; that may have benefits, if they succeed, for others without any heavyweight in their corner.

I'm beginning to see some common themes in the arguments for or against being able to copyright AI-created works, who is the creator of a work, and various analogies to past technologies and whether they produced original work.

Here's another article on the subject, again behind a paywall, arguing that the US copyright office's decision in the graphic novel was wrong. I don't agree with the point of view - I'm still stuck on the fact that none of the current generative AI tools would be able to produce if they hadn't ripped off masses of work to "train" themselves. The comments are lengthy and in many ways more interesting that the opinion piece The author also wrote �Creators Take Control: How NFTs Revolutionize Art, Business, and Entertainment.�. I haven't read that, but it doesn't surprise me that these two points of view come from the same person.

https://www.washingtonpost.com/opinions/2023/04/27/artificial-intelligence-copyright-decision-misguided/

In talking about the use of prompts as creative intellectual labor, he writes:

"If the creator doesn�t like the result, she can refine the prompts � as Kashtanova did hundreds of times � until the image matches her intellectual conception. The more one uses prompts, the better one becomes at predicting the results. That�s why new jobs as prompt engineers have sprouted. Businesses seek the fruits of these creators� �intellectual labor.�"

The derivative nature of this creativity reminds me of the furor over Shepard Fairey's Hope poster (although that case was complicated by the fact that he lied about the source image in a legal action)

The article includes the by-now-common analogy to photography and discusses an 1884 supreme court case about a photograph of Oscar Wilde where the court concluded that photographs were not simple mechanical reproductions, but �representatives of original intellectual conceptions of the author.� I think the fallacy is in comparing the system behind generative AI to the equipment that takes a photograph (and even more off base if you include post-processing software that the creator of a photograph almost always uses. If generative AI systems were based on robots having gone out to paint and photograph Yosemite, the deep dark woods, etc., there'd be some fair comparison, but that's not what has been done.

The top three most liked comments on the article:

"Sorry dude. Copyright Office is right. You're wrong. Besides your analogies fail. A jazz artist is wholly creating their improvisations. A prompt "artist" is tweaking someone else's. In music that is called arranging not creating. Arrangements dont get copyrights."

"Exactly. I'm an engineer who does computer coding. If I ask chatGPT to create a code to, say, run a basic spectral analysis on gappy satellite altimetry data, using a specific non-FFT method, I absolutely SHOULD be allowed to USE the results (assuming they're right, which they might well not be for technical reasons, but I digress). But I surely shouldn't be able to COPYRIGHT that code. Because I didn't write it. It's literally made from a computer recombination of the work of other previous coders and mathematicians. It's not "mine", even if I can use it for free.

There's nothing sacred about art that makes it different than anything else. If you want to copyright it, make it yourself. Otherwise, be grateful that you can use the skill of other people (AI programmers and real artists) to get your work done well for you."

"And this is generally true of the current machine learning trained on big data model of AI. Every sentence, every image, comes from recombining the words and images the AI is trained on, a huge portion of which was nominally protected by copyright. Current AI models are the worst enemy of copyright in history."

The current head of the US Federal Trade Commission wrote an op-ed piece for the New York Times (paywall) but there's a good summary in Ars Technica

https://arstechnica.com/information-technology/2023/05/we-must-regulate-ai-ftc-chair-khan-says/

There aren't very many comments in the NY Times - all comments are moderated, but even so only 267? An article about Tucker Carlson's texts had 2617 Comments are pretty negative about her approach - either because they are opposed to regulation in principle or because they don't see anything sufficiently specific about how to rein in tech companies. An example:

"There isn't a single specific suggestion in this article about how to regulate the use of AI.  It just lists the problems with which we are all already familiar.  If the head of a federal agency is going to place an article in the NYT, it should offer more than this."

There was one comment about content creators: "not a peep in this opinion about protection of intellectual property and creators work.   Just like streaming music, no doubt the AI lobbyists are on the job right now, making sure govt looks the other way as their clients appropriate the value of creative work."

This Wall Street Journal article (paywall) is about AI and advertising - along with the turmoil resulting from the shift from traditional TV to streaming and how the advertising dollars are following (or not). The reason I think this might have an impact on stock agencies is that a good chunk of our licenses are for advertising uses.

Digital Video Publishers Tout AI and New Metrics at NewFronts Sales Events
https://www.wsj.com/articles/digital-video-publishers-tout-ai-and-new-metrics-at-newfronts-sales-events-97a42c4f

A couple of interesting quotes:

"Players in free ad-supported television or ad-supported video on demand, smart-TV manufacturers like LG and Samsung, premium video publishers and others at the NewFronts all reported growing viewership and sought to convince buyers they deserved video ad spending..."


"As Writers Guild members marched outside of Peacock�s presentation, in part protesting studios� reluctance to regulate AI-generated material in television and movies, media executives sold AI as a technology to elevate buyers� advertising deals.
Roku described a new AI-driven feature that scans programming on its ad-supported Roku Channel to match advertisers, their products and their campaigns with relevant scenes in shows and movies.
�Imagine promoting a new phone line in the ad break after every time E.T. phones home, or advertising an apparel line every time Tim Gunn says �make it work� on �Project Runway,�� Julian Mintz, Roku�s head of U.S. brand sales, told ad buyers.""

Here is a 34 year old writer who lost his income because of Chatgpt

https://www.reddit.com/r/ChatGPT/comments/139o1q6/lost_all_my_content_writing_contracts_feeling/

I've just read it. One particular sentence struck me, and I couldn't resist commenting (on site) :

"For me, writing is like taking a sh*t: I don't have a choice."

We all do that, we just don't sell it.

Take it easy, plumber is quite a dignified profession. Perhaps more than selling words...

However I suspect that's just a prank: check his grammar, his spelling, his style, and then ask yourselves if anyone could really make a living out of that.

 This is a legal decision on what level of transformation constitutes fair use. As fair use is one of the arguments offered as to why generative AI images trained on copyrighted work is legally OK, I think it's worth adding to this list

https://petapixel.com/2023/05/18/supreme-court-rules-andy-warhols-prince-art-is-copyright-infringement/
https://www.hollywoodreporter.com/business/business-news/supreme-court-andy-warhol-prince-copyright-1235495647/
https://www.billboard.com/pro/andy-warhol-prince-supreme-court-copyright-case-ruling/
https://www.cooley.com/news/insight/2023/2023-05-18-supreme-court-rules-andy-warhols-prince-portraits-not-fair-use
https://www.theguardian.com/artanddesign/2023/may/18/andy-warhol-copyright-prince-paintings-lawsuit
https://variety.com/2023/music/news/andy-warhol-prince-images-supreme-court-copyright-1235617622/

i have wondered if it was written by Chatgtp as a prank, but thought there are too many mistakes....

you can ask  chatGPT to include grammatical mistakes, misspellings.

this was done in an experiment to see if teachers could detect which essays were ai generated - the teachers failed!

Quote from: cobalt on May 07, 2023, 21:21

i have wondered if it was written by Chatgtp as a prank, but thought there are too many mistakes....

you can ask  chatGPT to include grammatical mistakes, misspellings.

this was done in an experiment to see if teachers could detect which essays were ai generated - the teachers failed!

OMG!

Never thought about that possibility!

Thank you to everyone who keeps posting useful links here.

As things develop this will be our own little documentary for future stock historians�

Quote from: cascoly on May 19, 2023, 13:15

Quote from: cobalt on May 07, 2023, 21:21

i have wondered if it was written by Chatgtp as a prank, but thought there are too many mistakes....

you can ask  chatGPT to include grammatical mistakes, misspellings.

this was done in an experiment to see if teachers could detect which essays were ai generated - the teachers failed!

OMG!

Never thought about that possibility!

Thank you to everyone who keeps posting useful links here.

As things develop this will be our own little documentary for future stock historians�

on a positive note - teachers are inverting tasks - former in class tasks are done at home - short essays are written in class w/o electronics

Just a follow up to my last post in this topic...

The EU has come one-step closer to having the world's first law regulating Artificial Intelligence. The European Parliament voted today in favor to approve the AI Act. Next step, the talks will now begin with EU countries in the Council on the final form of the law. The aim is to reach an agreement by the end of this year.

From the series of measures, I highlight the following extracted from the news of the European Parliament:

"Generative AI, like ChatGPT, would have to comply with transparency requirements:

    Disclosing that the content was generated by AI
    Designing the model to prevent it from generating illegal content
    Publishing summaries of copyrighted data used for training

Limited risk
Limited risk AI systems should comply with minimal transparency requirements that would allow users to make informed decisions. After interacting with the applications, the user can then decide whether they want to continue using it. Users should be made aware when they are interacting with AI. This includes AI systems that generate or manipulate image, audio or video content, for example deepfakes."


link to European parliament news: https://www.europarl.europa.eu/news/en/headlines/society/20230601STO93804/eu-ai-act-first-regulation-on-artificial-intelligence

Quote from: Evaristo tenscadisto on June 14, 2023, 15:34

Just a follow up to my last post in this topic...

The EU has come one-step closer to having the world's first law regulating Artificial Intelligence. The European Parliament voted today in favor to approve the AI Act. Next step, the talks will now begin with EU countries in the Council on the final form of the law. The aim is to reach an agreement by the end of this year.

From the series of measures, I highlight the following extracted from the news of the European Parliament:

"Generative AI, like ChatGPT, would have to comply with transparency requirements:

    Disclosing that the content was generated by AI
    Designing the model to prevent it from generating illegal content
    Publishing summaries of copyrighted data used for training

Limited risk
Limited risk AI systems should comply with minimal transparency requirements that would allow users to make informed decisions. After interacting with the applications, the user can then decide whether they want to continue using it. Users should be made aware when they are interacting with AI. This includes AI systems that generate or manipulate image, audio or video content, for example deepfakes."


link to European parliament news: https://www.europarl.europa.eu/news/en/headlines/society/20230601STO93804/eu-ai-act-first-regulation-on-artificial-intelligence

Yes, but with all the ideas that the European parliament has, what impact will it make? Also this statement is way to vague. What actual laws do you think will come out of it? It will take years (if even something will get written black on white) and then it's already there and nothing can be done about it anymore. That will be the reality of it. Maybe they will fine some big American company or so by then (to collect some money) but everything will go on as usual.

Well... i really don't know at this point but I think we will have a note or a banner whenever we interact with an AI device. This will be for soon and not years. For Generative AI platforms like midjourney or stable diffusion probably they must provide some evidences where the work was based on. I don't see big issues at the moment since i see it as fair use or transformation art/work in my humble opinion.

But chat gpt is different since it can provide code and knowledge about things that goes beyond fair use. For example, in December last year there were people trying to find a workaround to ask Chatgpt how to rob a bank. Of course you can not ask Chatgpt directly, so they come up with the idea of roleplaying with chatgpt to extract the idea how to do it and what was the best way. 

Let me ask you how can Chatgpt have such knowledge? Where did AI learn this from? And if this goes to other fields like weapons, virus or terrorism?

That's why EU parliament wants chatgpt to Design the model to prevent it from generating illegal content and Publishing summaries of copyrighted data used for training to guarantee safety to all EU citizens. I don't believe this will take years and I wonder why CEO of ChatGPT developer OpenAI said "the firm may pull out of the Europe Union if current draft legislation is not toned down"? I wonder what Chatgpt have been learning... probably the good, the bad and the ugly. 

Quote from: SVH on June 14, 2023, 15:44

Yes, but with all the ideas that the European parliament has, what impact will it make?

I posted a news article about this law already where it stated that for example GPT Chat has announced that if the law  that the copyrighted source material that was used to train an AI has to be made public would pass, they would rather withdraw from the European market than do that, because they fear legal consequences. That would be a rather big impact, don't you think?

It might, but will it come this far? Will then ChatGPT dissapear in the EU? I wonder.

Quote from: Evaristo tenscadisto on June 14, 2023, 15:34

...

"Generative AI, like ChatGPT, would have to comply with transparency requirements:
...
    Publishing summaries of copyrighted data used for training


... This includes AI systems that generate or manipulate image, audio or video content, for example deepfakes."

...

a good first start - I'm hope they consider:

how can those offering ai know what images were used when the systems don't provide that info & how would it be saved &H displayed?  besides working thru agencies, artists offer their work on  independent website's & POD sites such as FAA

This is not a rule for microstock sites offering AI images for sale, this is a rule for companies that develope AI. They, so sites like ChatGTP or Midjourney or stable difusion would have to show which images they used to train their AI.