Statement on AI and AI-based projects
As generative artificial intelligence becomes more and more prevalent in creative projects, I wanted to write a short statement on my stance on AI, both in my personal projects, and in commercial commissions.
In its current state, I believe AI is unfit for handling any sort of serious responsibility, both on a technical and an ethical level. I do not implement AI in my personal projects, and I do not accept work on projects that are largely based on AI. I believe that the current trend of products which rely on ChatGPT or a similar service as a part of its primary feature set is unethical and unsustainable. Below I go into a bit more detail, but this is my final stance. If your project or project idea is largely based on AI, there is no need to contact me.
If you are interested in my reasoning for rejecting any AI-based projects, please read on. If you would like to discuss any of this with me, feel free to reach out via e-mail. My reasoning for not working on AI projects is as follows:
1. Results
While the technology behind generative AI, such as ChatGPT, Microsoft Copilot or Midjourney, is incredibly impressive on a theoretical level, the practical results often leave much to be desired. For worldbuilding and writing projects, ChatGPT often generates uninspired ideas and typically recognizable mediocre prose. For programming projects, AI services that generate code aren't able to take greater context into consideration, often generating code that refers to non-existing functions, misinterprets publicly available documentation, or explicitly just does not work. AI generated images are usually jarring, inconsistent in style, and instantly recognizable as AI, and therefore unfit for any sort of commercial use.
2. Sustainability
There are several major issues with the sustainability of using AI in commercial projects. On the small scale, code generated by AI, if functional, is still often low quality. Utilizing it leads to a codebase that very quickly accrues technical debt, and the potential time saved writing code is spent tenfold fixing bugs and issues that come up as a result of poor practices. On the large scale, if a project is built with its primary features requiring access to an AI API of any sorts, it is relying on the future availability and affordability of these services. While commercial access to, for example, the ChatGPT API is currently relatively cheap, it is hard to believe that it will stay this way. Building a product that heavily relies on such an API is commercially risky at best, entirely unsustainable at worst.
3. Ethics
Generative AI services have been widely criticized for their impact on employment, privacy, data security, and the environment. Corporations are replacing skilled writers and artists, relying on mediocre AI output to do their jobs. Data transparency in AI black-box systems is currently unachievable, and privacy is in no way guaranteed. Large language models have been proven to have a tremendeous negative impact on the environment, particularly in the training phase. Because of these reasons, I do not believe that heavy usage of generative AI in commercial projects is ethically justifiable.
Thanks for reading this article by Tadeas Jun! Have any thoughts, feedback, or just want to chat? Contact the author at contact@tadeasjun.com, or on Discord at @tadeasjun. Tip them on their PayPal or hire them as a world designer, writer, or software engineer via their Portfolio page.