Most chatbots inherently 'left-leaning,' but with capacity to 'learn' bias, study finds

Research has shown that many chatbots, including ChatGPT, exhibit a left-of-centre bias, which can be altered by training with politically aligned data. David Rozado's study achieved this, creating left, right, and neutral versions of chatbots by fine-tuning their algorithms with specific textual content, thereby influencing their political orientations.