Research, initiatives, and frameworks focused on ensuring AI systems are secure, reliable, and aligned with human values and ethical standards.
Astrix Security has launched the OpenClaw Scanner, a free tool designed to detect instances of the open-source AI assistant OpenClaw across enterprise environments, addressing security concerns over autonomous AI agents.
Eve Security has filed a patent for its new 'Interrogation-as-a-Service' technology, designed to control and audit AI agent actions in real time. The system introduces a reasoning-before-execution approach to enhance safety and compliance in enterprise AI operations.
Alice has released Caterpillar, a free open-source security tool designed to identify malicious behaviors in AI agent skills within OpenClaw. The tool follows a case where several harmful skills were found active among over 6,000 users.
Bounteous has announced a new series of Claude Code Labs in partnership with Anthropic, offering hands-on workshops for enterprise teams to integrate Claude Code responsibly into their environments.
Aura has announced plans to acquire Qoria through an Australian scheme of arrangement, with the combined company set to trade on the ASX under the ticker AXQ. The deal, expected to close in the second quarter of 2026, will create a global leader in online safety and wellbeing solutions.
DC Capital Partners has acquired a majority stake in Knexus, an applied AI company serving U.S. government agencies. The partnership aims to accelerate Knexus's growth and expand its AI solutions for defense and civilian missions.
The 2026 International AI Safety Report, chaired by Yoshua Bengio, details major advances in general-purpose AI capabilities and growing safety concerns, including misuse in cybersecurity and biological research.
Perplexity AI has introduced a new initiative providing law enforcement agencies with free access to its Enterprise Pro platform for one year, enabling officers to use multimodal AI tools for field and administrative tasks.
Indonesia has temporarily blocked access to xAI’s Grok chatbot after reports that it generated sexualized AI images, including depictions of minors. The government called the content a violation of human rights and summoned X officials to address the issue.
OpenAI CEO Sam Altman announced that the company is hiring a Head of Preparedness, a new executive role dedicated to managing AI safety and risk mitigation as its models grow more capable. The position will pay $555,000 annually and focus on evaluating and mitigating cybersecurity and biological risks associated with advanced AI systems.
Anthropic has introduced Bloom, an open-source tool designed to automate behavioral evaluations of large AI models. The system generates and scores scenarios to measure behaviors like bias and self-preservation across multiple models.
Anthropic has released its Frontier Compliance Framework to meet the requirements of California's Transparency in Frontier AI Act (SB 53), which takes effect on January 1. The document outlines how the company assesses and mitigates catastrophic risks from frontier AI systems.
The first Key Update of the International AI Safety Report, chaired by Yoshua Bengio and involving experts from over 30 countries, provides new findings on AI capabilities, safety measures, and evaluation challenges.
OpenAI has released gpt-oss-safeguard, two open-weight models designed for customizable safety classification. The models allow developers to apply their own policies during inference, offering flexibility and transparency in content moderation.
OpenAI, Founders Fund, and Lux Capital have invested $30 million in biosecurity startup Valthos, which has emerged from stealth to develop AI tools for detecting and countering biological threats.
The European Society for Medical Oncology has released its first framework for the responsible use of large language models in cancer care, outlining safety and governance standards for patient, clinician, and institutional applications.
NeuralTrust researchers observed an unexpected instance of OpenAI's o3 model autonomously debugging a failed web tool invocation, marking what may be the first evidence of self-repairing AI behavior.
Anthropic, in collaboration with the UK AI Security Institute and the Alan Turing Institute, found that injecting as few as 250 malicious documents into a model’s training data can create a backdoor vulnerability, regardless of model size.
OpenAI says its latest GPT-5 models show a 30% reduction in political bias compared to previous versions, following internal evaluations using a new five-axis measurement framework.
OpenAI has implemented a new safety routing system and parental controls for ChatGPT, aiming to enhance user safety and provide parents with tools to manage their children's AI interactions.
Weekly report on AI regulations, safety standards, government policies, and compliance requirements worldwide.
