Bias of AI-Generated Content: An Examination of News Produced by Large Language Models
This research addresses bias in AI-generated content, a critical issue for users and developers of LLMs, though it is incremental as it applies existing evaluation methods to new models.
The study investigated bias in AI-generated news content from seven large language models, including ChatGPT and LLaMA, by comparing it to original articles from The New York Times and Reuters, revealing substantial gender and racial biases, with ChatGPT showing the lowest bias and being the only model to decline generation with biased prompts.
Large language models (LLMs) have the potential to transform our lives and work through the content they generate, known as AI-Generated Content (AIGC). To harness this transformation, we need to understand the limitations of LLMs. Here, we investigate the bias of AIGC produced by seven representative LLMs, including ChatGPT and LLaMA. We collect news articles from The New York Times and Reuters, both known for their dedication to provide unbiased news. We then apply each examined LLM to generate news content with headlines of these news articles as prompts, and evaluate the gender and racial biases of the AIGC produced by the LLM by comparing the AIGC and the original news articles. We further analyze the gender bias of each LLM under biased prompts by adding gender-biased messages to prompts constructed from these news headlines. Our study reveals that the AIGC produced by each examined LLM demonstrates substantial gender and racial biases. Moreover, the AIGC generated by each LLM exhibits notable discrimination against females and individuals of the Black race. Among the LLMs, the AIGC generated by ChatGPT demonstrates the lowest level of bias, and ChatGPT is the sole model capable of declining content generation when provided with biased prompts.