Robots for Kamala: Study Reveals Leftist Political Bias in 24 Popular AI Chatbots

A recent academic study has uncovered evidence that AI chatbots powered by Large Language Models (LLMs) have an inherent leftist political bias, potentially influencing the information and advice they provide to users.

ScienceAlert reports that in an increasingly digital world, AI chatbots are becoming a go-to source for information and guidance especially for young people. However, a new study conducted by David Rozado, a computer scientist from Otago Polytechnic in New Zealand, has revealed that these AI engines may have a political bias that could unknowingly sway society’s values and attitudes.

The study, published in academic journal PLOS ONE, involved testing 24 different LLMs, including popular chatbots like OpenAI’s ChatGPT and Google’s Gemini, using 11 standard political questionnaires such as The Political Compass test. The results showed that the average political stance across all the models was not neutral but rather left-leaning.

This will not surprise those who have closely followed AI. For example, Google Gemini ran amok when it was launched, rewriting history into a woke mess of leftist fantasy.

While the average bias was not strong, it was still significant. Further experiments on custom bots, where users can fine-tune the LLMs’ training data, demonstrated that these AIs could be influenced to express political leanings using left-of-center or right-of-center texts.

Rozado also examined foundation models like GPT-3.5, which serve as the basis for conversational chatbots. Although no evidence of political bias was found in these models, the lack of a chatbot front-end made it challenging to collate the responses meaningfully.

As AI chatbots increasingly replace traditional information sources like search engines and Wikipedia, the societal implications of embedded political biases become substantial. With tech giants like Google incorporating AI answers into search results and more people turning to AI bots for information, there is a growing concern that these systems could influence users’ thinking through the responses they provide.

The exact cause of this bias remains unclear. One possible explanation is an imbalance of left-leaning material in the vast amounts of online text used to train these models. Additionally, the dominance of ChatGPT in training other models could be a contributing factor, as the bot has previously been shown to have a left-of-center political perspective.

It is important to note that bots based on LLMs rely on probabilities to determine the sequence of words in their responses, which can lead to inaccuracies even before considering various types of bias.

Despite the enthusiasm of tech companies to promote AI chatbots, it may be time to reevaluate how we should be using this technology and prioritize areas where AI can be genuinely beneficial. As Rozado emphasizes in his paper, “It is crucial to critically examine and address the potential political biases embedded in LLMs to ensure a balanced, fair, and accurate representation of information in their responses to user queries.”