Amidst the ongoing strike by the Writers Guild of America (WGA), the idea of utilizing artificial intelligence (AI) to replace human writers has ignited a fiery debate within the industry.
However, the concerns go beyond the question of who will be replaced.
Who's Replacing Who?
In a study co-authored by OpenAI, the creator of ChatGPT, it has been revealed that 80% of the US and UK workforce is expected to experience substantial impacts due to the emergence of large language models (LLMs).
The study predicts that about 19% of professionals will witness at least 50% of their work being exposed to ChatGPT and its various iterations. This is precisely what recent WGA demonstrations have been calling to prevent.
The WGA has expressed reservations regarding the use of AI in the writing process, advocating for AI models to be employed as tools rather than as creators of final products.
AI Tech Breaking Copyright Laws
One of the main concerns raised by the WGA is the issue of copyrighting AI-generated text, as writers fear the possibility of AI receiving credit or becoming the sole author of a work.
John August, a screenwriter and member of the WGA negotiating committee, voiced these concerns about AI, stating, "We don't want our material feeding them, and we also don't want to be fixing their sloppy first drafts."
Read also: Meta: New MTIA AI Chip Will Power the Company's Internal Workloads, Silicon Chip Better than CPUs
ChatGPT, a large language model trained on a vast corpus of internet text, raises questions about potential copyright infringements due to the lack of explicit consent for specific training data inclusion.
OpenAI, the company behind ChatGPT, fed the tool approximately 300 billion words systematically scraped from the internet. This includes personal information obtained without consent.
Moreover, reports have found that the scraped data used to train ChatGPT could contain proprietary or copyrighted material, with some instances even suggesting the model's ability to generate entire Harry Potter books.
Simply put, no online data is safe from being grabbed by the AI system.
Just an AI, Never a 'Source Material'
Reuters tells us that writers also worry that studios might utilize AI to create poorly written scripts and then hire them to fix them for lower pay.
To prevent this, a union proposed that AI-generated material should not be considered "literary material" or "source material" in writer contracts, ensuring that writers cannot be paid less to revise AI-generated scripts.
AI Writers Remain Limited
Gizmodo tells us that ChatGPT operates based on statistical patterns and lacks intention, a proper understanding of meaning, and the ability to refine its output. These limitations hinder its genuine writing capabilities.
The same report showcased an experiment conducted with ChatGPT yielded a script showcasing its strengths and limitations.
While the AI-generated text could serve as a useful reference for identifying pitfalls in writing, it fell short in critical areas such as style, emotional depth, coherent storytelling, and compelling dialogue or scenes.
It became apparent that the AI needed more intuition, creativity, and experience necessary for effective storytelling.
About the article image: A Bing Image Creator-produced image of the prompt "A robot and a human in an intense writing competition inside a boxing ring."
Stay posted here at Tech Times.
