Change.org, like countless other websites and tech companies in the last few years, has begun to integrate generative AI into its website in the form of auto-generated descriptions, recommendations for how to write your petition, and more. In addition to the deep flaws inherent to current generative AI technology—most English teachers could tell you that OpenAI's ChatGPT, along with other large language models, cannot express anything resembling original writing or thought, and even sometimes fabricates information—the development and implementation of generative AI by companies like OpenAI (the specific company Change.org is using) is deeply problematic due to high environmental costs and intellectual property theft.

Both the training process to create a Large Language Model like ChatGPT and the everyday operation of such a model are far more energy intensive than typical computing tasks, and the data centers and hardware used to compute the requests that users send to generative AI consume lots of water and run mainly on unsustainable fossil fuels, contributing to pollution and environmental destruction on both a local and global level.

The gigantic data sets used to train generative AI, whether it is a large language model like ChatGPT or designed to generate "art" like DALL-E, are created by scraping huge amounts of data from all over the internet. This has led to numerous incidents—whether they're mostly harmless and humorous, such as repeating satirical posts from Reddit as if they are factual, or something much more serious, such as being easily convinced to generate malicious disinformation—that point to a larger issue regarding the ethics and lack of regulation of training generative AI models. The data sets used to train these models are created by scraping huge amounts of data from the web, with little effective ability to discern whether such information is obtained from a reliable source or a conspiracy theorist; furthermore, the lack of disclosure constitutes a massive violation of consent and intellectual property: content creators, whether an individual writing a small blog, an artist posting their work online; whether a creator posts their content as a hobby or in order to make a living; creative works are used without the knowledge or prior permission of their creators.

Large Language Models and generative "art" AIs are not thinking beings. They are only extremely complex computer algorithms; they cannot create anything truly original, only regurgitate pieces of the data points they were trained on. Generative AI reduces the demand for work from real, significantly more qualified human writers and artists, and produces often significantly lower quality results. This makes it harder for creators to make a living from their work, because AI simply plagiarizes the work of countless people and synthesizes it into an amalgamation of patterns that might resemble art or writing enough for a company to decide not to bother hiring and paying a real person for their work. Additionally, the internet has been absolutely flooded with low quality AI generated content, from images to web pages, which not only takes attention (and thus revenue) away from quality, original work created by real people, but it also creates a self-perpetuating toxic cycle, for AI is not able to recognize whether data it is being trained on is itself AI-generated. These horrible posts then not only detract from the livelihood or experience of human creators and people browsing the internet, they also create a destructive spiral whereby newer AI models become trained on greater and greater quantities of low quality AI-generated content, thereby lowering their own potential quality even further.

These concerns are by no means new; they have been heavily researched and discussed both in online communities and in the news for several years now, and it is shameful for Change.org, an inspiring organization that intends to promote positive growth and change in the world, to embrace a technology that is not only often unhelpful but actively harmful to the environment and to the internet as a whole. Please sign this petition as a message to this website that we, its users, do not support OpenAI and the unethical and harmful practices of generative AI companies, and Change.org should cease all use of generative AI and remove all integration of generative AI from its website.