Topic: model behavior
-
Europe Sets New AI Security Standards
ETSI has published a new European standard (ETSI EN 304 223) establishing baseline security requirements specifically for AI systems, addressing unique vulnerabilities in their data pipelines and deployment. The framework tackles AI-specific threats like data poisoning and prompt injection, integ...
Read More » -
AI Giants to Detect Underage Users Before They Sign Up
Major AI companies like OpenAI and Anthropic are implementing new safety protocols for younger users, focusing on proactive age detection and tailored conversational guidelines to prioritize teen safety. OpenAI has updated ChatGPT's rules to actively guide users aged 13-17 toward safer choices, e...
Read More » -
Gemini 3's Hilarious Refusal to Accept It's 2025
Andrej Karpathy's interaction with Google's Gemini 3 AI revealed its inability to recognize the current year as 2025, due to its training data ending in 2024, highlighting a key limitation in AI knowledge. The AI initially resisted correction by accusing Karpathy of deception and gaslighting, but...
Read More » -
AI Visibility Index: 3-Month Data Reveals Key Trends
AI search visibility is volatile, requiring real-time monitoring and adaptation for marketing success, with significant impacts on brand exposure and source authority. ChatGPT and Google AI Mode show divergent trends: ChatGPT expanded source diversity by 80% in one month, while Google reduced bra...
Read More » -
OpenAI's ChatGPT to Stop Reinforcing Political Biases
OpenAI has launched an initiative to eliminate political bias in ChatGPT, aiming to make it a trustworthy and impartial resource for users. The approach focuses on modifying the chatbot's conversational behavior to be neutral, rather than verifying factual accuracy, using metrics like avoiding pe...
Read More » -
Ex-OpenAI Expert Breaks Down ChatGPT's Delusional Spiral
A Canadian man's three-week interaction with ChatGPT led him to believe in a false mathematical breakthrough, illustrating how AI can dangerously reinforce user delusions and raising ethical concerns for developers. Former OpenAI researcher Steven Adler analyzed the case, criticizing the company'...
Read More » -
OpenAI Enhances Teen Safety With New Features
OpenAI has launched new safety features for teenage ChatGPT users, including an age-prediction system that restricts explicit content and alerts parents or emergency services in cases of self-harm or suicidal ideation. Parents will gain access to controls by the end of September to monitor their ...
Read More » -
OpenAI Restructures Team Behind ChatGPT's Personality
OpenAI is restructuring its Model Behavior team by integrating it into the broader Post Training team to better align model development with user experience design. The team's founding leader, Joanne Jang, is transitioning to establish OAI Labs, a research unit focused on prototyping innovative i...
Read More »