The 'truth serum' for AI: OpenAI’s new method for training models to confess their mistakes

OpenAI and NextDC Ltd. have announced a partnership to develop a large-scale data center in Sydney, marking a significant step in enhancing data infrastructure in Australia. This collaboration aims to support the growing demand for AI technologies and services, particularly as OpenAI continues to expand its offerings.

Anthropic president Daniela Amodei has expressed confidence that the market will ultimately reward safe artificial intelligence (AI), countering the Trump administration's view that regulation stifles the industry. Amodei's perspective highlights a belief in the potential for responsible AI development to thrive despite regulatory challenges.

OpenAI is facing intensified competition from Google, particularly with the rapid rise of Google's Gemini 3, which has gained 200 million users in just three months. In response, OpenAI CEO Sam Altman has declared a 'code red' for ChatGPT, emphasizing the urgent need for improvements to maintain its market position.

Amazon Web Services (AWS) has launched Kiro powers, a new system designed to enhance AI coding assistants by providing them with specialized knowledge on demand, rather than loading all capabilities at once. This innovation was announced at the re:Invent conference in Las Vegas and aims to address inefficiencies in current AI coding tools.

A recent study by Gong reveals that sales teams utilizing artificial intelligence (AI) generate 77% more revenue per representative, indicating a significant shift in how revenue leaders perceive AI in business decision-making. The study analyzed over 7.1 million sales opportunities across more than 3,600 companies and surveyed over 3,000 global revenue leaders from various countries, including the US, UK, Australia, and Germany.

OpenAI has been ordered to provide ChatGPT data as part of a new court ruling, raising concerns about user privacy and the handling of AI-generated conversations. This ruling could set a significant precedent for how data from AI interactions is managed and disclosed in legal contexts.

A research team from China and Hong Kong has introduced a new memory architecture called General Agentic Memory (GAM) aimed at addressing the issue of 'context rot' in AI models, which leads to the loss of information during lengthy interactions. This dual-agent system separates memory functions to enhance information retention and retrieval, potentially improving the performance of AI assistants in complex tasks.

Anthropic and OpenAI have recently showcased their respective AI models, Claude Opus 4.5 and GPT-5, highlighting their distinct approaches to security validation through system cards and red-team exercises. Anthropic's extensive 153-page system card contrasts with OpenAI's 60-page version, revealing differing methodologies in assessing AI robustness and security metrics.