By now, you’ve likely tried out one of the new AI-based image generation tools, which ‘sample’ a range of image repository websites and online references to create all new visuals based on text prompts.
DALL·E is the most well-known of these new apps, while Midjourney has also become popular in recent months, enabling users to create some startling visual artworks, with virtually no effort at all.
But what are your usage rights to the visuals you create – and for marketers, can you actually use these images in your content, without potential copyright concerns?
Right now, it seems that you can – though there are some provisos to consider.
“Subject to your compliance with these terms and our Content Policy, you may use Generations for any legal purpose, including for commercial use. This means you may sell your rights to the Generations you create, incorporate them into works such as books, websites, and presentations, and otherwise commercialize them.”
Yes, you can even sell the visuals you create, though most stock photo platforms are now re-assessing whether they’ll actually accept such for sale.
This week, Getty Images became the latest platform to ban the upload and sale of illustrations generated through AI art tools, which, according to Getty, is due to:
“…concerns with respect to the copyright of outputs from these models and unaddressed rights issues with respect to the imagery, the image metadata and those individuals contained within the imagery.”
Part of the concern here is that the visuals that are used as the source material for these AI generated depictions may not be licensed for commercial use.
Though even that’s not necessarily a definitive legal barrier.
As explained by The Verge:
“Software like Stable Diffusion [another AI art tool] is trained on copyrighted images scraped from the web, including personal art blogs, news sites, and stock photo sites like Getty Images. The act of scraping is legal in the US, and it seems the output of the software is covered by “fair use” doctrine. But fair use provides weaker protection to commercial activity like selling pictures, and some artists whose work has been scraped and imitated by companies making AI image generators have called for new laws to regulate this domain.”
Indeed, various proposals have been put forward to potentially regulate and even restrict the use of these tools to protect artists, many of whom could well be out of the job as a result. But any such rules are not in place as yet, and it could take years before a legal consensus is established as to how to better protect artists whose work is sourced in the back-end.
There are even questions over the technical process of creation, and how that applies to legal protection in this sense. Back in February, the US Copyright Office effectively implied that AI-generated images can’t be copyrighted at all as an element of ‘human authorship’ is required.
In terms of specific content policies, DALL·E’s usage terms state that people cannot use the app to ‘create, upload, or share images that are not G-rated or that could cause harm’.
So no depictions of violence or hate symbols, while the DALL·E team also encourages users to proactively disclose AI involvement in their content.
DALL·E’s additional guidelines are:
- Do not upload images of people without their consent.
- Do not upload images to which you do not hold appropriate usage rights.
- Do not create images of public figures.
This is where further complications could come in. As noted by JumpStory, users of AI image generation tools should be wary of potential copyright concerns when looking to create images that include real people, as they may end up pulling in pictures of people’s actual faces.
Again, the legal specifics here are complex, and really, there’s no true precedent to go on, so how such a case might actually be prosecuted is unclear. But if you are looking to generate images of people, there may be complications, if that visual ends up directly resembling an actual person.
Clearly stating that the image is AI-generated will, in most cases, provide some level of clarity. But as a precautionary measure, avoiding clear depictions of people’s faces in your created images could be a safer bet.
MidJourney’s terms also make it clear violations of intellectual property are not acceptable:
“If you knowingly infringe someone else’s intellectual property, and that costs us money, we’re going to come find you and collect that money from you. We might also do other stuff, like try to get a court to make you pay our attorney’s fees. Don’t do it.”
Oddly tough talk for legal documentation, but the impetus is clear – while you can use these tools to create art, creating clearly derivative or IP infringing images could be problematic. User discretion, in this sense, is advised.
But really, that’s where things stand, from a legal perspective – while these systems take elements from other visuals online, the actual image that you’ve created has never existed till you created it, and is therefore not subject to copyright because your prompt is, in effect, the original source.
At some stage, the legal technicalities around such may change – and I do suspect, at some time, somebody will hold an AI art show or similar, or sell a collection of AI-generated art online which depicts significant elements of other artists’ work, and that will spark a new legal debate over what constitutes intellectual property violation in this respect.
But right now, full use of the images created in these tools is largely fine, as per the terms stated in the documentation of the tools themselves.
Note: This is not legal advice, and it’s worth checking with your own legal team to clarify your company’s stance on such before going ahead.