‘I made this, so I own it’

I joined the user group for an AI image SaaS that I am tinkering with to see what the fuss is about and to get a feel for this ‘prompt engineering’ thing. The objective of the group is for people to share what they have generated with the prompts and settings they used so others can learn how to use the software and reproduce the results.

Amongst some very neat generations of different types of scenes, I have seen several attempts at images of celebrities and protected IP with questions about the legality of using these images for commercial gain.

Many of the answers from other users indicated that, as the image creator, the copyright belongs to them and therefore have the right to do whatever they want with them.

The questions about rights, ownership, and legality that keep coming up in these sorts of forums are why we need to keep talking about this and not just invoke cancel culture on the technology.

If you have opted out of (actively) using AI on principle, you should still consider not opting out of the conversation altogether, because ultimately you already are subjected to AI and it’s going to become even more ubiquitous in the future. Being involved in these conversations is in your best interest. There are also those who are opting into the technology who don’t know, understand, or care about some of the very valid reasons behind many of those principles, and discourse is what grows knowledge on both sides so that we can see what is coming in this burgeoning space and hopefully mitigate crossing too much over to the Dark Side(TM). While regulation in other areas has mixed results, we can at least be part of the process of determining what is or isn’t acceptable use in our society or industry, and help pave the way for best practices and standards that can support the way to legislate or regulate down the road.

(Mis)representation

Let’s return to the example of the generated image of a celebrity (or any other real human being).

There is a difference between a real photo you have taken of a person doing something they have actually done (and the consent issues there) and a picture you have constructed to which the person has not consented, even if it is depicting something they might naturally do.

There is also a difference between something that is represented as, or is, a photograph, and other forms of artistic expression. Irrespective of how that bit of expression is created, how are you representing the work or your subject and for what purpose or intent?

In terms of legalities, consider the distinction between “freedom of speech” and “freedom of the press” or their equivalents in your country and how it applies to the difference between, for example, a sculpture that makes a political statement versus a photo of a real-life political event.

Laws relating to copyright, right of publicity, etc. will also address things like whether that expression represents the artist’s skill or style, as some sort of derivative work or interpretation, or a representation of fact or news.

You don’t need to have a lawyer’s understanding of the minutiae, but it’s important to understand that there is nuance attached to what these are and recognise that they are different things for the time when you are crafting your expression or interpretation.

Also, if you are using AI as part of your work product, what are you saying you are doing and what is the scope?

As a basic example, if you’re an impressionist who does a great Michael Caine, are you selling your work as an impressionist of Michael Caine or Michael Caine himself? Are those images of your dad who looks like Tom Hanks being represented as your dad who looks like Tom Hanks or Tom Hanks himself?

Now, if an impressionist or doppelganger decides to train their own AI using their own voice or visual representation to generate impressions or images, what do you call that work product? Is it still you doing an impression, or being depicted as a doppelganger, or is it an AI generation of you doing those things?

Will someone pay you the same for a generation of your skill if they know it’s not actually you? If you choose not to disclose that the product is AI generated and it comes to light that it is, what is the impact on you and the industry? Will the industry be accepting of expressions of art of this kind? Will it cause artists in this space to be devalued or create more space for exploitation? Do you think it is ethical? Will your clients?

Look at what was being raised as part of the SAG-AFTRA strikes and the major studios asking background actors to be scanned so their digital replicas can be used as they see fit in some cases without consent or limit on scope. Some authors are already affected by publishers wanting to train AI on their work:

What precedent does your work and how you’re representing it set, and is it the kind of thing that will only affect you, or could it affect your industry¹?

Copyright

Also, how do we protect our work product when AI is mixed in? What is the copyright in AI? The prompt? It’s ‘just’ a string of words, so where is the line of copyright here? The license to use the software? No. The algorithm or configuration of the AI engine implemented? That does not belong to the end user. The data we feed it? Well, that’s why we have the mess with ChatGPT and similar platforms that we do. The content has been farmed from a variety of sources that could be a bit murky in the consent area.

As it is, a federal judge in the US ruled that AI-generated art cannot be copyrighted, so that’s going to put a halt to some of that².

Another interesting exchange that came up in these AI image forums was when a user shared a prompt they used to create an image and another user used the prompt verbatim to generate a different set of images and stated an intention to sell the results. The first user was upset by this and was informed by the community that they should only share prompts they are happy for others to use³. I digress.

‘But we make derivative content all the time!’ I hear you say. ‘What about fan art?’ I hear you ask.

Fan created content is still bound by the copyright of the owners of the original IP. This is also a fun issue, look it up, heaps of precedent here⁴⁵ and financial compensation for fan works is a problem area. Usually leaning towards the “not permitted” side of things. What you create though, even if it’s derivative, is still yours. That creative expression belongs to you, but that doesn’t mean you can sell it.

And so many of these types of arguments show how little people understand about copyright, ownership when it comes to creative expression, and ‘public domain’.

Where to from here?

There is a fear that AI will render creative artists obsolete. While as a tech/science person, I cannot definitively say that it’s outside the realm of possibility, I feel we’re still quite a way away from that. But that doesn’t change the fact that we need to talk about what we as artists and creators deem is and isn’t acceptable with how AI fits into our respective industries, how our image and work product is created and used, and how we can protect our works, or at least be compensated appropriately for it⁶.

Long before this kind of thing was a concern, Karen Jacobsen became the the voice of millions of GPS units. She has one of the most recognisable voices in the world, but at the time, contracts and licensing didn’t account for this breadth of usage, so despite Karen’s voice being literally everywhere, she had only received a one-time payment, which was probably a sore point when her family was suffering financial hardship. I highly recommend you listen to the episode of Twenty Thousand Hertz Recalculating, featuring Karen Jacobsen and the voice of Polish Google Maps, Jarosław Juszkiewicz.

Just like with other parts of our lives, we can also set boundaries. Samuel L Jackson has been de-aged and granted permission for his image to be used in this way for these specific projects and ensures there is no “in perpetuity” type clause for any of these contracts. It is his image which he has consented for the studios to use for specific use cases only.

So, when it comes to you and your work, how would you like to see your industry unfold as AI inevitably enters the fray? Even if you don’t intend to use it directly, if you did or were required to, how would you use it in an ethical and responsible way, and what protections or boundaries would you like to have on your self and your work product?

It’s something to think about even if you don’t call yourself an artist or a creator, because the scope for applied AI is only limited by the power of human imagination which has a reputation for being both great and terrible.

---

STEAM Powered

It really has been a minute. I've been rationing out my energy as I recover and STEAM Powered kept powering on, but my writing did not.

Helen McKenzie is a cartographer, data visualiser, and geospatial advocate who is passionate about how data and making information beautiful can give us more meaning to the way we live in the world around us.

April Wensel is the founder of Compassionate Coding, helping companies and tech professionals communicate more and effectively by putting the humanity back into technology.

Laura Langdon is a developer advocate who has the pleasure of combining her explorations and experience in theatre, computer science, mathematics, education, and data science into a role that rolls all of that into one perfect package.

Evelyn Lee is Head of Workplace Strategy and Innovation at Slack Technologies, and Founder of Practice of Architecture, and speaks about the future of the architecture industry and how we apply systems thinking in both physical and organisational environments.

Sirisha Kuchimanchi is an entrepreneur, speaker, and former engineering and manufacturing executive. We speak about knowing your personal worth, and it's not just about money.

Dr Sandy Chong is a founding member of the Sustainable Development Goals Forum in WA, former president of the United Nations Association of WA, and member of the World Economic Forum Expert Network. and speaks about how UN Sustainable Development Goals benefit us all, and are for now, not later.

Eleonora Moratto is the Biology Ballerina. A passionate SciArtist, Eleonora wants people to embrace polymathy and cultivate multiple careers that can support us at different stages of our lives.

Dr Lydia Pethick is a veterinarian, motivational speaker, television presenter, and advocate for mental health and well-being in the veterinary industry, raising awareness of the challenges of being a vet, and how the community can better support each other.

Quite Interesting

In case you were wondering, Experimental replication shows knives manufactured from frozen human feces do not work, and Potatoes are better than human blood for making space concrete bricks.

Cora Harrington wrote a great thread about ‘clothes rich’ and the economics of textiles historically, with some context about the Lendbreen Tunic, which dates back to the early first millennium AD. #TextilesNerd

On an entirely different note, it's very much not the season for this, but I made the Best Thick Hot Chocolate Recipe (Italian Style) from Healthy Little Cravings the other day, and it's delicious.

---

Thanks for reading, and see you next time!

Stay curious,

— Michele

---

Footnotes

1. Are you in it for a quick buck, or are you planning to be here for the long haul? ↩

2. I’m now wondering if this might impact algorithmically composed works or things like Demoscene. ↩

3. Something something, what’s good for the goose, something. ↩

4. I refer you to Anne Rice and how problematic fanfic writing is to some authors. There’s a Kotaku article about some of the related incidents here. You can’t realistically stop the fandoms, but bear in mind many writers and authors will not read fanfic about their own properties and ask fans to not send them any story ideas so that they can continue producing their works without risk of legal impact. ↩

5. I cannot tell you how much I laugh/cried when AMC launched what was effectively a fanfic platform and encouraging Interview with the Vampire fans to create content. ↩

6. I wrote a piece about the need for discourse in the area of new technologies previously. ↩