All articles tagged with #ai censorship
Sony has patented an AI-based censorship tool aimed at making games more child-friendly by automatically detecting and censoring inappropriate content on PS5, PS6, Xbox, and Nintendo consoles, allowing users to customize content filtering and potentially making mature games accessible to all ages.
A digital consultant warns that Google Gemini's AI model is biased due to parameters defining "toxicity" and shaping what information remains "invisible." The model's training data and filters have been criticized for being ideologically skewed, leading to problematic outputs. The concern lies in the lack of transparency regarding how toxicity is defined and how it shapes the digital world. Google CEO Sundar Pichai acknowledged the issue and plans to fix Gemini's bias, but concerns persist about the impact on users and the company's position in the AI market.
Nvidia has released its open-source "NeMo Guardrails" software, which acts as a censorship bot for apps powered by large language models. The software is designed to help chatbots stay on topic and prevent them from spewing misinformation, offering toxic or outright racist responses, or creating malicious code. NeMo works on top of older language models like OpenAI’s GPT-3 and Google’s T5, and is supposed to work with "all major LLMs supported by LangChain, including OpenAI’s GPT-4." However, it remains unclear how effective an open-source guardrail might be in preventing AI from lying or cheating.