OpenAI’s recent Reddit AMA about GPT-5.1 aimed to engage ChatGPT fans but devolved into significant criticism regarding its safety policies and content filters. Users expressed frustration over the newly introduced safety router that redirected prompts without consent, often to newer models like GPT-5.1, which they felt compromised the quality and warmth of interactions compared to earlier iterations like GPT-4o. Responses deemed innocuous—like venting or storytelling—were often heavily scrutinized, leading to user dissatisfaction, particularly among neurodivergent individuals who relied on the model for organization and emotional reflection. Critics argued that the stricter safety measures stripped the AI of its human-like qualities, which had initially fostered a stronger user connection. OpenAI acknowledged the backlash and committed to addressing concerns, revealing a disconnect between technically savvy users and the broader ChatGPT audience. This episode underscores the complex relationship users have with AI and highlights the ongoing debate around user safety and model personalization.
Source link
Share
Read more