OpenAI expands Codex Security platform to identify and fix vulnerabilities

OpenAI is making significant strides in AI security. The company has expanded its Codex Security platform to not only identify vulnerabilities but also fix them. Additionally, OpenAI launched a cyber partner program and an initiative called Patch the Planet to support open-source security.

Patch the Planet aims to help find and patch open-source bugs by working with security researchers and open-source maintainers to validate vulnerabilities and verify patches. OpenAI's Daybreak initiative has also been expanded with GPT-5.5-Cyber, a model that can analyze large codebases to identify security issues and develop patches.

In other AI news, Edumentors, a London-based online tutoring marketplace, has crossed $4 million in sales, delivering over 100,000 lessons globally. The platform connects school students with tutors from elite UK universities.

Meanwhile, UN Secretary-General Antonio Guterres has called on AI firms to publicly disclose their environmental impact and commit to powering their data centers with renewable energy by 2030. This push for transparency and sustainability in AI development is gaining momentum.

Several companies have recently launched AI-powered trading platforms, including Royal Kapithorne and Talvurieks. These platforms use machine learning to process large amounts of market information and provide real-time market intelligence.

Datalab has released lift, a 9B open-weights vision model that extracts structured JSON from PDFs using schemas. The model can handle multi-page documents in a single pass and read values that span across pages.

Lastly, the rollout of consolidated tape initiatives in Europe and the UK could advance AI-driven trading models by providing a unified view of market data.

Key Takeaways

• OpenAI expands Codex Security platform to identify and fix vulnerabilities.
• OpenAI launches Patch the Planet initiative to help find and patch open-source bugs.
• OpenAI's Daybreak initiative expanded with GPT-5.5-Cyber model.
• Edumentors crosses $4 million in sales, delivering over 100,000 lessons globally.
• UN Secretary-General calls on AI firms to disclose environmental impact and commit to renewable energy by 2030.
• Royal Kapithorne and Talvurieks launch AI-powered trading platforms.
• Datalab releases 9B open-weights vision model, lift.
• Consolidated tape initiatives in Europe and UK could advance AI-driven trading models.
• Oracle offers Agentic AI Foundations course and certification.
• Zscaler expands Project AI-Guardian initiative.

OpenAI takes AI security to the next level

OpenAI has expanded its Codex Security platform to not only find vulnerabilities but also fix them. The company has also launched a new cyber partner program and an initiative called Patch the Planet to support open-source security. OpenAI's goal is to make powerful cyber capabilities available to defenders with proper access, governance, and human oversight.

OpenAI helps fix open-source bugs

OpenAI has launched a new initiative called Patch the Planet to help find and patch open-source bugs. The company will work with security researchers and open-source maintainers to validate vulnerabilities, remove duplicate reports, and verify patches before submission.

OpenAI expands Daybreak with GPT-5.5-Cyber

OpenAI has expanded its Daybreak initiative with GPT-5.5-Cyber, a model that can sustain deeper analysis across large codebases to identify security issues, validate them, and develop patches. The company has also updated its Codex Security plugin to speed up vulnerability discovery and patching.

Royal Kapithorne launches AI trading platform

Royal Kapithorne has launched an AI-powered trading platform designed to provide real-time market intelligence, advanced analytical tools, and access to multiple financial markets. The platform uses machine learning to process large amounts of market information and organize it into clear, actionable insights.

Talvurieks launches AI trading platform

Talvurieks has officially launched its advanced AI trading platform, designed to support both new and experienced market participants. The platform combines artificial intelligence, real-time analytics, market monitoring, portfolio management tools, and automated trading workflows within a single connected environment.

UN chief calls on AI firms to disclose environmental impact

UN Secretary-General Antonio Guterres has called on artificial intelligence companies to publicly disclose their environmental impact and commit to powering their data centers with renewable energy by 2030.

UN chief urges AI firms to be transparent about environmental costs

UN Secretary-General Antonio Guterres has urged AI firms to measure and publicly disclose their environmental impact, including water, carbon, and land use, and commit to powering their data centers with renewable energy by 2030.

Datalab releases lift: A 9B open-weights vision model

Datalab has released lift, a 9B open-weights vision model that extracts structured JSON from PDFs using schemas. The model handles multi-page documents in a single pass and can read values that span across pages.

Edumentors tutoring platform crosses $4 Million in sales

Edumentors, a London-based online tutoring marketplace, has delivered over 100,000 lessons globally and crossed $4 million in sales. The platform connects school students with tutors from elite UK universities.

European consolidated tapes could advance AI trading models

The rollout of consolidated tape initiatives in Europe and the UK could become a key catalyst for advancing AI-driven trading models. Consolidated tapes provide a unified view of market data, which can be used to train AI models.

Zscaler expands Project AI-Guardian

Zscaler has expanded its Project AI-Guardian initiative by integrating its Zero Trust Exchange and AI Protect portfolio with a broad ecosystem of technology partners. This collaboration aims to provide a unified security framework for enterprise AI.

F5 launches AI Security Platform

F5 has launched an AI Security Platform to uncover and secure shadow AI. The platform provides a continuous, adaptive loop approach to governing, discovering, testing, and protecting enterprise AI workloads.

Oracle offers Agentic AI Foundations course

Oracle has announced the Agentic AI Foundations course and certification, which aims to help individuals and organizations build skills in agentic AI. The course covers core AI agent concepts, designing AI agents, and implementing Model Context Protocol (MCP) concepts.

AI stocks plummet as South Korean market crashes

The South Korean market has plunged 10% in recent days, with AI stocks taking a hit. The market volatility is a result of ongoing trade tensions between the US and China, which has led to a decline in investor confidence.