Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Adobe Firefly: AI Art Generator

adobe.com

Post reply on HN
Print view View original

311–320 of 496 posts

Re: Adobe Firefly: AI Art Generator

#311

Post by smrtinsert »

lelandfe wrote: post #50

From that page's FAQs: > trained on a dataset of Adobe Stock, along with openly licensed work and public domain content where copyright has expired > We do not train on any Creative Cloud subscribers’ personal content. For Adobe Stock contributors, the content is part of the Firefly training dataset, in accordance with Stock Contributor license agreements. The first model did not train on Behance. Not sure what "firs…

I would not be surprised if behind the scenes they are starting the lobbying engine to safely mine whatever they want. The universe of existing content out there is simply too enticing and out there. This is Google Search vs authors all over again.

Re: Adobe Firefly: AI Art Generator

#312

Post by balls187 »

airstrike wrote: post #266
balls187 wrote: post #249

Earlier quoted context omitted.

This runs into the core problem with technology--we answer "What can we do" before "Should we do it" and "What are the impacts" Let's say you take your hello kitty dot art, and make a poster promoting a commercial event. You then take it to FedEx Kinkos and use a self-service copy machine to make 1000 copies. You could reasonable argue that you are violation of copyright infringement, and the photocopier / FedEx kink…

Whoever put the poster up for display and reaped some reward out of it is liable for damages. Everyone else is just doing their job in the supply chain. We want supply chains to work for the good of the economy, which is a proxy for increasing availability and reducing prices of "goods and services" to the average person.

Imagine a paying Adobe CC customer.

They use Firefly to generate a poster, and unbeknownst to them, the image it generated is a reasonable facsimile of a copyrighted/trademark character.

The person has inadvertently committed copyright infringement.

So does Firefly need to come with a warning?

The safer solution, to the chagrin of another commenter, is for Adobe to neuter the tool by only training on data in which Adobe has express permission to use.

Re: Adobe Firefly: AI Art Generator

#314

Post by BornInChicago »

egypturnash wrote: post #160
judge2020 wrote: post #53

Earlier quoted context omitted.

Still on the fence for whether or not you should be able to opt out of training (I'm sure many artists would love to "opt out" of humans looking at their art if the human intends to, or might, copy the artists' style at some point).

hi, I'm an artist, I do not give a shit about other humans looking at my work, I am delighted when a younger pro comes to me and thanks me for what they learnt from my work. That tells me they were fascinated enough with it to look at it and analyze it again and again. I made a connection with them via my drawing skills. I am catastrophically unhappy at the prospect of a corporation ingesting a copy of my work and st…

I'm an artist as well. I think this can happen whenever anyone sees your art anywhere online. They can copy it. They won't tell you about it. They might copy it really well. And they might copy not just your technical style, but what your art says and how it says it.

Re: Adobe Firefly: AI Art Generator

#315

Post by BornInChicago »

waboremo wrote: post #143
grondo4 wrote: post #135

Earlier quoted context omitted.

> i.e. the act of reading may cause accidental infringement when implementing the "things you learn" Surely you know this isn't the case right? Maybe you're confused because we're talking about programming and not a different creative artform? Great artists read, watch and consume copyrighted works of art all day, if they didn't they wouldn't be great artists. And yet the content they produce is entirely there own, f…

Artists get into trouble all the time for producing works very close to something that already exist. That's like the number one reason artists get shunned in the communities they were in.

I would seriously question if this happens all the time, these days. The whole copyright thing is way behind the digital and internet revolution. Look at what the Prince case did for transformation copyright fair use.

Re: Adobe Firefly: AI Art Generator

#316

Post by brucethemoose2 »

int_19h wrote: post #276
brucethemoose2 wrote: post #34

Earlier quoted context omitted.

These models can be run locally, but I see far too many businesses and bloggers resign themselves to renting OpenAI's cloud instead of tinkering with LLaMA/Alpaca, the community Stable Diffusion finetunes and such. Adobe does have a legal edge here, which is interesting and perhaps actually worth a subscription if needed.

The problem is that GPT-3.5 is so cheap, and the results that you get out of it are still quite a bit better than LLaMA/Alpaca. There just doesn't seem to be any solid economic reason to run it locally except to keep inputs and/or outputs private.

That is not sustainable though. OpenAI is in the "burning money to gain market share" phase.

Re: Adobe Firefly: AI Art Generator

#317

Post by klabb3 »

mesh wrote: post #139
Filligree wrote: post #120

Earlier quoted context omitted.

One of the great aspects of open-source stable diffusion (civitai.com et al.) is there's a model for every purpose. Does your inpainting model work with every style? Or is it going to have trouble matching the content for e.g. specific fanart?

It would have trouble matching on trademarked styles, or individual artists / creators styles. One of the primary goals for Firefly is to provide a model that can generate output that respects creators and is safe for commercial use. (I work for Adobe)

I understand the intent but the result will clearly sway in the direction of protecting big brands, artists, and individual styles. There’s simply no way that it couldn’t. At some point in the pipeline, there’s a blocklist for copyrighted works of a finite size that’s decided by employees, no?

Re: Adobe Firefly: AI Art Generator

#318

Post by wwweston »

krisoft wrote: post #201
egypturnash wrote: post #149

Earlier quoted context omitted.

I am definitely glad to see attention being paid to ethical sourcing of training images but I am curious: did the people who made all those stock images get paid for their work being used for training? Did they check a box that explicitly said "Adobe can train an AI on my work"? Or is there a little clause lurking in the Adobe Stock agreement that says this can be done without even a single purchase happening?

I'm not a lawyer and I don't work for Adobe. :) The contributor agreement linked from here[1] is this: [2] "You grant us a non-exclusive, worldwide, perpetual, fully-paid, and royalty-free license to use, reproduce, publicly display, publicly perform, distribute, index, translate, and modify the Work for the purposes of operating the Website; presenting, distributing, marketing, promoting, and licensing the Work to u…

I'm starting to think that use of works in a training set is a category not covered well by existing copyright law, and it may be important to require separate explicit opt-in agreement by law (and receipt of some consideration in return) in order to be considered legitimate use.

The vast majority of copyrighted works were conceived and negotiated under conditions where ML reproduction capabilities didn't exist and nobody knew what related value they were negotiating away or buying.

Re: Adobe Firefly: AI Art Generator

#319

Post by sebzim4500 »

adam_arthur wrote: post #211
dopeboy wrote: post #11

What this reinforces is that unlike with previous big innovations (cloud, iphone, etc), incumbents will not sit on their laurels with the AI wave. They are aggressively integrating it into their products which (1) provides a relatively cheap step function upgrade and (2) keeps the barrier high for startups to use AI as their wedge. I attribute the speed at which incumbents are integrating AI into their products to a…

Creating AI models has proven to simply be easier than other past innovations. Much lower barrier to entry, the knowledge seems to be spread pervasively within months of breakthroughs. People seem to take offense at this idea, but the proof is in the pudding. Every week there's a new company with a new model coming out. What good did Google's "AI Talent" do for them when OpenAI leapfrogged them with only a few hundre…

That's not really true though. 4 months on and no one else is close to matching the original ChatGPT.

It's too early to say how hard this is, for all we know no one but OpenAI will match it before 2024.

Re: Adobe Firefly: AI Art Generator

#320

Post by usrusr »

TylerE wrote: post #179
usrusr wrote: post #73

Earlier quoted context omitted.

Kids at street corners have idea and vision, no shortage in that. What matters is putting together idea and vision with execution. And for anything that is beyond basement scope, execution further subdivides into craft and access. Access to the means, and that is true for Hollywood as much as it is true for the smallest-time painter who might not be good at making friends with gallerists. The outliers are those that…

Kids on corners have ideas. Very few of them are GOOD ideas. Anyone can say 'Make a movie with lots of aliens and lasers'. It still takes Ridley Scott to make Alien .

Yet at the same time that "Pinback chased by the beach ball monster" scene that eventually evolved into Alien is hilariously deep in "kids on street corner ideas" territory
Post reply on HN