26 Oct Graphic Artists Guild Issues Statement of Concern with AI Image Generators
The Graphic Artists Guild has issued a statement of concern to the IFRRO General Assembly on AI image generators, and the detrimental impact this technology is having on graphic artists. The Guild is conducting an thorough overview of the practices and technology behind AI generators, how visual artists are using this technology, IP confusion and concerns, and the impact on the labor market for visual artists. We are also partnering with other creators’ organizations to better understand how AI technologies are affecting the livelihoods of all authors, including musicians, composers/songwriters, writers, and photographers.
Our statement is as follows:
GRAPHIC ARTISTS GUILD CONCERNS WITH AI IMAGE GENERATORS
Artificial intelligence image generators can provide a great benefit to visual artists. For example, AI image generators are used as a tool for ideation and inspiration, and visual artists will be hired for jobs requiring the curatorial skills required to effectively use AI image generators. However, there are ethical and legal concerns raised by the use of AI image generators, and the consequences for visual artists will be grave.
Unauthorized Use of Creators’ Imagery
The neural networks driving AI generators have been trained from existing imagery. Some AI generators scrape the Internet for imagery, and others rely on images which have been submitted by users. The licensing process visual artists rely on is simply ignored. Overall, there is no consideration for the copyrights or licensing of those images. Some AI image platforms provide a means for artists to have their work removed from the image base, creating yet another realm in which artists must invest time and effort to police the use of their work. Additionally, once a work has been used for neural training, the damage has already been done; with millions of users on AI image generating platforms daily, thousands of images trained from that artwork may have already been generated.
Confusion on Copyrights
The US Copyright Office has stated that AI generated images created solely by artificial intelligence cannot be copyrighted. However, the Office has recognized and generated a registration certificate for AI generated images where human creativity – in determining the text prompts to generate the image, in curating the results, and in altering the result – has played a role. This has resulted in confusion in interpreting whether a work is an original work of human authorship or not. Additionally, where AI generated images result in a work similar to an original work, unresolved questions of fair use are raised. This has created an environment in which the original creators of visual imagery are unsure of their protections under copyright law. Infringers are emboldened to submit copyrighted works to AI image generators, and indeed to claim copyright to the result.
Ethical Concerns in the Imitation of Artists’ Works
The text prompts used to generate AI imagery permit users to request images which ape the distinctive style of individual visual artists. This will enable works which appear to have been created by a recognizable artist to be generated and licensed, bypassing commissioning or licensing work from that artist. This raises grave ethical concerns which have yet to be addressed by the technology sector invested in AI image generators. Once again, as with previous technological advancements which have facilitated
infringement, visual artists who have spent years and effort honing their craft are the fodder to create considerable income, without remuneration, for others.
Work Displacement
We recognize that visual artists – in particular graphic artists such as designers and illustrators – natively have the prompting and curating skills required to be effective users of AI image generators. This will open up job possibilities and more effective ways of self-generating work for visual artists. However, it is inarguable that AI image generators will reduce the market for visual artists. Those same prompting and curating skills can be learned by individuals with no interest or ability in creating imagery without the intervention of AI image generators. Young artists will be discouraged from investing in educational programs in traditional media. The overall loss in job opportunities and the resultant reduction in skilled artists cannot be overlooked.
There Can Be a Better Path Forward
The Graphic Artists Guild recognizes that AI image generating technology is the wave of the future, and that this technology presents opportunities as well as challenges to visual artists. This technology is still in the early phases of its evolution, hence the lack of clarity on rights issues. However, we believe that this is an opportunity for the tech sector to take steps to recognize the damage this technology can do to the livelihood of visual artists and to propose standards and take measures to prevent the abuse of creators’ works:
- Permit visual artists to block the use of their works in the training of the AI image generator neural networks — for example, by leveraging image metadata
- Prepare ethical standards for users which recognize the illegitimacy of using AI image generators to bypass the hiring of visual artists by generating work which copies their distinctive style and ideation
- Provide robust tools for visual artists to police and remove once and for all any works which already exist in the databases of imagery used for neural training
Many of the AI image generating platforms already openly discuss the ethical considerations of using their technology to generate violent, racist, defamatory, or misleading/ deep fake imagery. We would like that conversation expanded to recognize the negative impact AI image generating platforms will have on visual artists. Visual artists have already suffered a grievous loss in licensing and commission income from past technological developments which facilitated the open theft of their work, such as file sharing, social media, e-commerce platforms, and the like. This is an opportunity for the technology sector to get ahead of the damage the AI image generating platforms will do to visual artists, and to recognize and address their impact on the creative community.