Art and "Generative AI" are NOT Equal
By Arte es Ética
Audiovisual narration
The year 2022 saw us overcome covid-19, but brought with it another type of expanding virality that affects thousands of people and the world’s economy: the so-called generative Artificial Intelligence (AI), a technology to which we are all exposed, as well as pressured to accept and normalize its use, often involuntarily, because we are constantly subjected to an advertising discourse that tries to convince us that its use is inevitable and even desirable.
Generative Artificial Intelligence is a field within AI, focused on the automatic production of data, text, images, and other content, based on the introduction of a brief text command (prompt) that tells the software what to generate and in what form. It’s important to refer to the results of AI as just that, as results or products extracted from human creative processes with which it is trained to generate automated derivations, so it is not about creations or works, terms that are proper of a human activity subject to copyright laws.
Generative AI has been presented by technology companies as „good news“, as a positive advance for society, but they do not mention the damage it is causing in different areas. One of them is related to its contribution to the deterioration of the environment and the aggravation of climate change because this type of technological models require a colossal amount of data that considerably increases the consumption of energy and natural resources needed to operate. Furthermore, this technology has been developed and trained with illegally obtained data, and its use facilitates the emergence of a series of problems that can irreversibly affect people’s lives.
Data mining or data scraping is the technique of taking millions of data from the Internet that governments allow some entities to do for research purposes. Universities or large laboratories dedicated to research, for example, in the area of mental or physical health, perform these activities, but they are obliged to anonymize the data in the processing of the extracted information. This modality of work is key and mandatory: keeping the identity of the people whose sensitive data is taken anonymous is an essential requirement in order not to harm their chances of getting a job or a health benefit, for example.
However, Stable Diffusion and its programmers have made public dozens of catalogs with the names of artists (anyone can find them on the Internet), with the premise that their technical „emulation“ is possible thanks to the technology they promote and the application they give to it.
Let’s review: Emad Mostaque founded Stability.AI at the same time as LAION, an alleged „non-profit “ charged with capturing nearly 6 billion images without consent, credit or compensation to the people who created those images or are featured in them.
That list of image + text pairs was then „ceded“ by LAION to Stability.AI to be used as a database in the training of its Stable Diffusion software, which, although released as open source, continues to pay multi-million dividends to the company as of August 2022 and constitutes the most important data laundering of the last decade.
This immense artistic heritage of the whole society was expropriated through the development of the Stable Diffusion model. This explains the leap in quality that this technology took in 2022 and is exploited to this day by all the platforms that rely on its dataset, such as Midjourney and Dall-E, among others.
Several of these models of generative AIs are promoted and operate using the names of the plundered artists and use them as part of the text commands to instruct the software to use machine learning of their works in the generation of outputs. This constitutes another instance of violation: the usurpation of the identity of the artists and of the moral and patrimonial rights that correspond to them.
The development of generative AI, both under the legal cover of nonprofit, claiming research purposes, and that which takes place using terms and conditions of very broad meaning, leaves those who work in the creative field in a situation of legal defenselessness that is very difficult to reverse. This is cruelly clear in the case of Adobe Stock users whose work was used to train Adobe Firefly, insofar as their acceptance of the contracts designed by the companies puts them in a position of responsibility for what happens to them.
The situation has the appearance of revictimization, being the result of a careful design of contracts by the legal teams of the companies, and the frequent lack of legal knowledge or availability of legal advice by those working in the creative field.
The more precarious the worker’s employment situation, the less likely he/she will be to seek advice to defend his/her rights or to avoid incurring in breaches that may harm him/her.
Holding the user accountable is one of the ways that generative AI development companies find to evade their responsibility and the law regarding explicit consent.
A regulation is necessary that clearly establishes that companies have the obligation to actively and explicitly inform collaborating users of any change in the policy of use of their content. It is also important to establish that they cannot be used to train generative technology until there is explicit consent from the collaborating users. Otherwise, the alienation of the product of creative work and the mechanization of artistic work could become a fact enshrined in jurisprudence that validates procedures such as Adobe’s, and this could then be extrapolated to all other vulnerable labor sectors.
Naturalizing this practice of holding the user exclusively responsible constitutes a precedent of serious asymmetry that, if legitimized, could be „the rule“ for making decisions in all similar cases in the future.
Generative artificial intelligence is not as innocent and harmless as its marketing campaign claims.
Instilling erroneous or inaccurate beliefs as universal axioms has become the most effective resource to convince everyone about the goodness of generative AI. One of the most repetitive is that generative AI is a „tool“ that facilitates the creative process of artists, when what it does is to override it and replace it with a series of written commands (prompts) intended, not to make the user’s imagination come true, but to locate in its database the most accurate result to the order received. A database crammed with millions of illicitly obtained copyrighted works.
Technology companies have fabricated the need and urgency to automate „tedious“ and „monotonous“ tasks, among which, it seems, is the creative process. They pose as saviors of sorts, as if the creative industry has asked for their help to give up their profession so that they can devote themselves exclusively to the contemplative life.
There is a notorious malicious will in distorting reality to adjust it to greedy purposes, appropriating the expression „democratization of art“ to ratify their feigned solidarity behavior.
With the excessive imposition of this type of technology in all the daily activities of the population, they try to implant a message of resignation and submission in those who are reluctant to give up their rights and freedoms in favor of the financial benefits of a handful of corporations.
Generative AIs operate today without any regulation whatsoever, and far from being mere tools for people applied to the creative field, they are presented as the ideal means for their replacement.
The scenario we are describing is related to labor problems because it affects our sector. But if we extrapolate this scenario of automatic content generation to the problem of information and information security, the danger that algorithms continue to take our data and generate similes of the instruments through which we inform ourselves, directly threatens any democratic order.
So is identity theft at a personal level through voice cloning for criminal purposes, such as virtual kidnapping or blackmail. Websites, newspaper articles, audios and videos are susceptible to being falsified by generative AIs and implanted in the „bubble“ of each user on social networks or skewing the circulation of information through channels copied by algorithms.
The whole society becomes an involuntary data worker under threat because they use the age-old method of instilling fear.
As citizens and workers, the advance of generative technology puts us more and more in the need to carefully read the „fine print“ of the contract of each digital product we use, discuss them and refrain from accepting them without an assurance that such terms and conditions are not implying assignment of inalienable rights to companies.
It also obliges us to group together, participate and propose the regulations that we need to be enshrined in law in defense of our sector, given that the speed of technological development is not closely followed by the legislative action of the States. The role of the latter, both in the legislative and educational aspects, generating not only legal bodies but also accessible information and providing advice to its citizens, is crucial.
Faced with a machine that completely replaces the process with an immediate and controlled result, how will young people overcome the frustration of learning? How will they elaborate their doubts, when the answer is provided to their partner in three seconds?
In what field of work will the great artists of tomorrow be formed, if the instances of simple commissions are solved with AIs? How will we sustain the attention, the infinite dedication required to build critical capacity to complexify our way of doing and its discursive resources, in the face of the expelling immediacy of an artificial competitor?
If everything that can be digitally recorded-the voice, the speech, the image of a person and the work product of his or her own body, everything, can be used as input to be replicated, derived or transformed, we are witnessing the end of authorial work. And, what is worse, the forced withdrawal of truth as a use value, while society is pushed to inhabit a fake life, fed with real experiences of people under virtual arrest: everything they say or do can and will be used against them.
Mental health is key to ensure the continuity of the members of the creative field. Chronic anxiety, anguish and frustration take their toll on the capacities of the artistic community, which is deprived of the value of its work, while countless sociopaths who glorify this software send them messages of hate and threats. When an algorithm takes control of the cultural industries, displacing the entire creative community, it makes the quality of its processes and results precarious and, appropriating jobs, accelerates everything to the point of chronic anxiety, reduces costs, and thus also the wealth of the entire society. Wealth that rightfully belongs to the human community.
What society would benefit from replacing millions of creators with an algorithm emulating their capabilities? No human society.
If, as active members of society, we do not assert our rights through government institutions, this scenario will continue to worsen, in which the devalue of what is „remixed“ by machines is imposed over the original value of what is created by people.
Authorships are today totally defenseless against automated plagiarism, since everything that is put into circulation on the Internet by its authors or anyone else is susceptible to being captured and subsumed as a training database for new versions of generative AI: a technology that is an instrument for the covert theft and replacement of workers; a technology developed for the sole purpose of enabling companies to dispense with their workers in order to „save costs“ and maximize their economic performance.
From Arte es Ética, we oppose data theft, privacy and intellectual property violations, the generation of audiovisual fakes and deepfakes, the exploitation of creatives, the proliferation of discriminatory biases, automation and job substitution.
We believe that it is urgent that an organized union of creatives from all audiovisual disciplines is needed to develop a set of ethical rules to ensure the healthy continuity of the creative profession, to dismantle the parasitic nature of this new industry by restoring the rule of law of the people
and strengthening regulatory frameworks on data mining, to prevent the development of new models of generative AI that compete unfairly, without adequately compensating their legitimate authors or creators. To achieve all this, we created Arte es Ética as a space to bring together Spanish-speaking artists and creatives and to promote collective actions of awareness and mediation with political representatives from Latin America and Spain.
Text used
- Manifiesto Arte es Ética
- Autoría y Autoencoders, de Santiago Caruso para Arte es Ética
- El “efecto Firefly”, de Luz Castro para Arte es Ética
- Desvirtuar la realidad , de Nia Soler para Arte es Ética
- DECÁLOGO DE LA FALSIFICACIÓN Y SUSTITUCIÓN DE VALOR HUMANO, de Santiago Caruso
y Naida Ochoa, para Arte es Ética - Creadores en contra de la IA: Piden se les pague por obras usadas sin permiso. Luis
Demano con Surya Palacios | ALTONIVEL
About the Author
Arte es Ética
Arte es Ética is a non-profit collective, which emerged at the end of 2022 as a result of the irruption of generative AI in our artistic fields and is integrated by creative professionals and independent visual artists (illustrators, designers, photographers, animators, painters, 3D modelers, among others) whose main objective is to achieve an ethical regulation of generative AI in Latin America and Spain. In addition, we work jointly with associations of other artistic disciplines, unions and cultural spaces in Latin America. Our main work is focused on informing about how generative AI affects the creative sector and bringing this information to the political representatives of the different Spanish-speaking countries.
Some Voices
Atilio Gambedotti, from Spain. Profesional Comic book artist/writer and illustrator
Axel Uriel González Pérez, from México. Concept and Visual Artist. Art Teacher.
Naida Jazmin Ochoa, from Argentina. Profesional illustrator and Graphic Designer
Nia Soler, from Spain. Editorial illustrator, painter and writer.
Some texts and illustrations are from
Luz Castro, from Argentina. Psychologist, writer, artist and Teacher.
Santiago Caruso, from Argentina. Visual Artist, writer and Art Teacher.
Logo Animation
Julián Medina, from Colombia. Filmmaker and Visual Artist.
Axel Uriel González Pérez, from México. Concept and Visual Artist. Art Teacher.