Lesley Thiel opinion on AI generated images
Tinha acabado de publicar O caos das imagens geradas por AI e o algorítmo indicou-me este texto escrito pela pintora irlandesa Lesley Thiel. Já era tarde e não pude escrever um comentário ao mesmo. A opinião dela reforça a minha ideia de que o uso destes geradores está desgovernado. Alguma coisa devia ser feita, o quê, exactamente, não sei. Enquanto isso, temos os "artistas" que publicam imagens AI para obter "atenção", ou o que for, e os "escritores" que escrevem palermices acerca do seu uso exactamente para o mesmo efeito, que, de igual maneira, se tornam imensamente populares. Bombar postagens apalhaçadas para manter o algorítmo entretido é uma tendência forte. Os textos questionadores, como o desta artista, que nos fazem parar e pensar no assunto, são relegados para segundo plano. É a vida de todos os dias nas redes sociais. Cada vez tenho menos paciência para ela.
Esta artista rejeita a utilização de modelos AI em qualquer fase do seu trabalho. E então eis o absurdo. Escreve ela: "Foi-me dito, por alguém que está a tentar proteger os artistas face ao ataque da IA, que a minha pintura ‘And We Praise Her’ foi sinalizada, por um software de deteção, como provavelmente gerada pela Stable Diffusion. Isto apesar de ter sido pintado vários anos antes de a geração de imagens de IA estar disponível."
Ela não considera que estas imagens possam ser chamadas de "arte". E acrescenta ainda: "Aqueles que usam geradores de imagens de IA geralmente afirmam que há imensa habilidade em definir os prompts necessários para criar imagens. Se um cliente viesse até mim e fornecesse uma descrição muito detalhada do que deseja incluir em uma encomenda, os direitos autorais da obra de arte resultante não seriam dele, mas sim meus. Não importa quantos ajustes e mudanças eles sugeriram. Eu sou o artista e a lei reconhece isso. Por que? Porque ainda é minha visão e minha habilidade que traduz essas sugestões em uma obra de arte física. Portanto, alguém que “cria” uma imagem de IA está a chegar às mentes de dezenas de milhares de artistas e a roubá-los, para que um computador cuspa algo que possam reivindicar como sendo o seu próprio trabalho."
E na linha do que escrevi na minha postagem anterior, ela escreve: "Neste momento, as imagens de IA estão inundando a Internet. As redes sociais estão cheias de contas que fazem passar imagens geradas por IA como arte e não admitem o engano, a batota. Isso pode não parecer prejudicial para o observador casual, mas está causando danos reais aos artistas digitais e também prejudicará os artistas tradicionais. As pessoas terão cada vez mais dificuldade em encontrar a arte desses criativos, quando ela estiver enterrada sob uma avalanche de imagens geradas por IA. E terão dificuldade em discernir qual arte é real e qual é falsa. E com isso, os artistas perderão um canal para a sua voz criativa: uma voz que a sociedade precisa ouvir, porque os artistas proporcionam uma visão do mundo que muitas vezes está fora do mundano e do dia a dia. "
Por fim, acrescenta ainda preocupações de ordem ambiental com a produção destas imagens: "No momento, os geradores de IA são em sua maioria de uso gratuito. Mas isso não significa que eles não tenham custo. Eles exigem grandes quantidades de poder de processamento. Estima-se que, até 2026, 30% da utilização de energia na Irlanda será destinada aos centros de dados que circundam Dublin, devido ao aumento da utilização de IA."
Por fim, acrescenta ainda preocupações de ordem ambiental com a produção destas imagens: "No momento, os geradores de IA são em sua maioria de uso gratuito. Mas isso não significa que eles não tenham custo. Eles exigem grandes quantidades de poder de processamento. Estima-se que, até 2026, 30% da utilização de energia na Irlanda será destinada aos centros de dados que circundam Dublin, devido ao aumento da utilização de IA."
E pergunta ainda pelas reais intenções destas empresas de pesquisa.
I don’t use AI generators at any point in the creation of my paintings. I don’t use them to come up with ideas, or to show me how to knit together the elements of a painting. I have never, and will never use an AI generator in my artistic practice. The concepts, successes and failings in the work are mine, and mine alone. I’m saying this because I was told, by someone who is attempting to protect artists in the face of the AI onslaught, that my painting ‘And We Praise Her’ was flagged, by a detection software, as probably generated by Stable Diffusion. This despite the fact that it was painted several years before AI image generation became available.
I’ve made my views on AI generated imagery pretty clear, in the past. I have decided not to call it ‘art’ because, in my view, that term applies solely to something created by the human mind and hand, be it with tools or a stylus. The images created by applications like Stable Diffusion, Midjourney, and Dall-E are not art, they are the stolen work of tens of thousands of artists, thrown into an algorithm pot and used to regurgitate something that a user thinks will gain them attention and clicks.
Those who use AI image generators usually claim that there is immense skill in defining the prompts needed to create images. If a client came to me and supplied a very detailed description of what they wanted to be included in a commission, the copyright of the resulting piece of art would not be theirs, it would be mine. No matter how many tweaks and changes they suggested. I am the artist and the law recognizes that. Why? Because it is still my vision and my ability that translates those prompts into a physical piece of art. So someone ‘creating’ an AI image is reaching into the minds of tens of thousands of artists and stealing from them, in order to have a computer spit out something they can claim as their own work.
Most artists continue their practice, no matter the struggle. It takes years to build up skills and find openings in the art world. For the few who are lucky enough to do well, there are thousands who struggle to even recoup their investment in their creative vision. But they keep going, through all the setbacks, because there is something inside them that needs to express itself through their art. It’s their way of speaking to the world.
Right now, AI images are flooding the internet. Social media platforms are full of accounts passing off AI generated images as their art, and failing to own up to the cheat. This may not seem harmful, to the casual observer, but it’s doing real damage to digital artists, and it’s also going to hurt traditional artists. People will find it more and more difficult to find the art of those creatives, when it’s buried under an avalanche of AI generated imagery. And they will find it hard to discern which art is real, and which is fake. And with that, the artists will lose a channel for their creative voice: a voice society needs to hear, because artists provide a view of the world that is often outside of the mundane and day to day. They speak to people’s hearts through their work. A soulless machine can never do that. It can never truly speak to you of shared experience and suffering; of love and hope and inspiration. It can only vomit back what it thinks you want to see, like a psychopath who knows how to imitate real human emotion.
There is another concern with all the AI generated imagery appearing on our feeds. It’s the cost. Right now, AI generators are mostly free to use. But that doesn’t mean they have no cost. They require huge amounts of processing power. It’s estimated that, by 2026, 30% of Ireland’s energy usage will be for the data centres that ring Dublin, because of the rise in AI usage. According to the IAEA, the amount of energy these technologies use is expected to explode. That’s energy that has to be generated in the real world and that is almost certainly created, at least in part, by the burning of fossil fuels. What will happen when we need more energy to cool homes in climate change generated heat waves? Or grow food indoors because our climate no longer allows it to thrive outside? What happens when there is a conflict between the survival of local populations and the profits of technology companies? If past experience is anything to go by, it will not end well for the people.
I have one final thought, in this admittedly long diatribe. What’s the end game for the people investing in AI generated imagery? Right now, they’re giving usage away, or asking for relatively small subscriptions. But these applications involve billions, if not trillions, in investment. Why? What is the ultimate purpose of their activities? Not their currently stated aims, but the reason why they would spend the GDP of a small country on developing something that steals human creativity.
I’ve made my views on AI generated imagery pretty clear, in the past. I have decided not to call it ‘art’ because, in my view, that term applies solely to something created by the human mind and hand, be it with tools or a stylus. The images created by applications like Stable Diffusion, Midjourney, and Dall-E are not art, they are the stolen work of tens of thousands of artists, thrown into an algorithm pot and used to regurgitate something that a user thinks will gain them attention and clicks.
Those who use AI image generators usually claim that there is immense skill in defining the prompts needed to create images. If a client came to me and supplied a very detailed description of what they wanted to be included in a commission, the copyright of the resulting piece of art would not be theirs, it would be mine. No matter how many tweaks and changes they suggested. I am the artist and the law recognizes that. Why? Because it is still my vision and my ability that translates those prompts into a physical piece of art. So someone ‘creating’ an AI image is reaching into the minds of tens of thousands of artists and stealing from them, in order to have a computer spit out something they can claim as their own work.
Most artists continue their practice, no matter the struggle. It takes years to build up skills and find openings in the art world. For the few who are lucky enough to do well, there are thousands who struggle to even recoup their investment in their creative vision. But they keep going, through all the setbacks, because there is something inside them that needs to express itself through their art. It’s their way of speaking to the world.
Right now, AI images are flooding the internet. Social media platforms are full of accounts passing off AI generated images as their art, and failing to own up to the cheat. This may not seem harmful, to the casual observer, but it’s doing real damage to digital artists, and it’s also going to hurt traditional artists. People will find it more and more difficult to find the art of those creatives, when it’s buried under an avalanche of AI generated imagery. And they will find it hard to discern which art is real, and which is fake. And with that, the artists will lose a channel for their creative voice: a voice society needs to hear, because artists provide a view of the world that is often outside of the mundane and day to day. They speak to people’s hearts through their work. A soulless machine can never do that. It can never truly speak to you of shared experience and suffering; of love and hope and inspiration. It can only vomit back what it thinks you want to see, like a psychopath who knows how to imitate real human emotion.
There is another concern with all the AI generated imagery appearing on our feeds. It’s the cost. Right now, AI generators are mostly free to use. But that doesn’t mean they have no cost. They require huge amounts of processing power. It’s estimated that, by 2026, 30% of Ireland’s energy usage will be for the data centres that ring Dublin, because of the rise in AI usage. According to the IAEA, the amount of energy these technologies use is expected to explode. That’s energy that has to be generated in the real world and that is almost certainly created, at least in part, by the burning of fossil fuels. What will happen when we need more energy to cool homes in climate change generated heat waves? Or grow food indoors because our climate no longer allows it to thrive outside? What happens when there is a conflict between the survival of local populations and the profits of technology companies? If past experience is anything to go by, it will not end well for the people.
I have one final thought, in this admittedly long diatribe. What’s the end game for the people investing in AI generated imagery? Right now, they’re giving usage away, or asking for relatively small subscriptions. But these applications involve billions, if not trillions, in investment. Why? What is the ultimate purpose of their activities? Not their currently stated aims, but the reason why they would spend the GDP of a small country on developing something that steals human creativity.
Fonte: https://www.facebook.com/lesley.thiel
Comentários